Snapshots of Radical Enzyme Catalysis: The Crystal

Snapshots of Radical Enzyme Catalysis: The Crystal
Structures of Biotin Synthase and Lysine 5,6-Aminomutase.
By
Frederick Berkovitch
B.S. Biochemistry
Brandeis University, Waltham, MA
Submitted to the Department of Chemistry in
Partial Fulfillment of the Requirements for the Degree of
DOCTOR OF PHILOSOPHY
In Biological Chemistry
At the
Massachusetts Institute of Technology
September 2004
© 2004 Massachusetts Institute of Technology
All rights reserved
/
Signature of Author
Z
I
--
r
-
,
-
I .
Department of Chemistry
August 2004
Certified by
Catherine L. Drennan
Associate Professor
Thesis supervisor
Accepted by
MASSACHUSS INSTUTE
OF TECHNOLOGY
SEP
15
2004
LIBRARIES
Robert W. Field
Chairman, Departmental Committee on Graduate Students
'ARCHIVES
This doctoral thesis has been examined by a Committee of the Department of Chemistry
as follows:
Professor JoAnne Stubbe
;~ ,
-
-
-_- --
·
.
....
I..1[.........,,i
Committee Chairperson
Novartis Professor of Chemistry
Professor of Biology
I
Professor Stephen J. Lippard
Arthur Amos Noyes Chlair- hdPrjssor
of Chemistry
Professor Catherine L. Drennan
Thesis supervisor
Associate Professor of Chemistry
-2-
To my family and to Shaunna
-3-
Snapshots of Radical Enzyme Catalysis: The Crystal
Structures of Biotin Synthase and Lysine 5,6-aminomutase
by
Frederick Berkovitch
Submitted to the Department of Chemistry in September 2004 in Partial
Fulfillment of the Requirements for the Degree of Doctor of Philosophy
Abstract
"AdoMet radical" enzymes utilize a 4Fe-4S cluster and S-adenosylmethionine, in
conjunction with a reducing system, to catalyze reactions that proceed through organic
radical intermediates. Genomics studies suggest that at least 600 such enzymes exist
across the archaea, bacteria, and eukaryotes, though few have been purified and
characterized. AdoMet radical enzymes participate in several biomedically important
processes, such as DNA biosynthesis and repair, cofactor biosynthesis, and bacterial
pathogenesis. To gain insight into this enzyme family, the crystal structure biotin
synthase from Escherichia coli, in complex with its substrate dethiobiotin, was
determined to 3.4 A resolution. Biotin synthase catalyzes a remarkable reaction, the
insertion of sulfur into dethiobiotin to form biotin (vitamin B8). The structure of biotin
synthase indicates that adenosylmethionine is coordinated to a unique Fe of the 4Fe-4S
cluster, in agreement with spectroscopic investigations from other laboratories.
Additionally, the structure reveals that the active site, located in the core of an (a/)
8
barrel, contains a unique 2Fe-2S cluster, in addition to the 4Fe-4S cluster,
adenosylmethionine, and dethiobiotin. Multiple lines of evidence suggest that in the
AdoMet radical enzymes, adenosylmethionine is a source of 5'-deoxyadenosyl radical, an
intermediate that is historically associated with the adenosylcobalamin-dependent radical
enzymes. Interestingly, some of the adenosylcobalamin radical enzymes catalyze
reactions that are very similar to those of AdoMet radical enzymes. One such example is
adenosylcobalamin-dependent lysine 5,6-aminomutase, which is analogous to lysine 2,3aminomutase, an AdoMet radical enzyme that is found in the same pathway. To further
probe the relationship of the adenosylmethionine and adenosylcobalamin-dependent
radical enzymes, the crystal structure of lysine 5,6-aminomutase was determined to 2.8 A
resolution. Like biotin synthase, the structure of lysine-5,6-aminomutase adheres to the
general theme of barrels or barrel-like structures for enzymes that utilize 5'deoxyadenosyl radicals in catalysis.
Additionally, the structure of lysine-5,6aminomutase implicates pyridoxal phosphate in a novel role, acting as a lock to prevent
the formation of aberrant radicals in the absence of substrate.
Thesis supervisor:
Title:
Catherine L. Drennan
Associate professor
-4-
I thank Professors Drennan, Jarrett, Frey, Licht, Klibanov, Lippard and Stubbe, and their
lab members, for enabling this work and for their counsel.
My brother Mike inspired me to go the distance, and continues to be a bottomless well of
wisdom and experience. My mother and father have always reminded me of what my goal was,
and had the utmost confidence in me. I always hoped that my success in graduate school would
inspire my little cousins Joe, Brian, and Jenny to set and achieve their own goals. Their parents,
my aunt and uncle, expected nothing less than success from me. My grandparents and greatgrandparents lived a life of unimaginable hardship and sacrifice in order to give me, my brother,
and my cousins the opportunity to live better lives. Through these last two years, I could always
turn to Shaunna when things were difficult. She is the kindest, most beautiful person I have ever
known. I don't know what I'd do without her. She took care of me during these last months made me food, told me so when I needed to sleep... My experience at MIT was enriched by the
company of the friends I've made. I thank them (you know who you are) for helping me,
teaching me, and sharing their experiences with me. I guess the most poignant thing I learned in
grad school is something that was so beautifully expressed by Edward Estlin Cummings:
O sweet spontaneous
earth how often have
the
doting
fingersof
prurient philosophers pinched
and
poked
thee
,has the naughty thumb
of science prodded
thy
beauty
.how
often have religions taken
thee upon their scraggy knees
squeezing and
buffeting thee that thou mightest conceive
gods
(but
true
to the incomparable
couch of death thy
rhythmic
lover
thou answerest
them only with
spring)
-5-
Table of Contents
Abstract .........................................
Acknowledgements ........
.....................
...............................................................
......................................
Table of Contents............................................................
.............................................................
List of Tables.................
..............
List of Figures............................................................
4
5
5...............
6
9
10
Chapter 1: Introduction to the structure of AdoMet and AdoCbl -dependent
radical enzymes
Abbreviations....................................................................................................... 13
1.1 AdoCbl-dependent
enzymes................................................................................
14
16
1.2 AdoMet radical enzymes............................................................
17
1.3 Why two radical cofactors? ............................................................
1.4 The diversity of 5 '-deoxyadenosyl radical enzymes............................................ 20
1.5 AdoCbl and AdoMet -dependent radical enzymes: structural information ......... 21
1.5.1
1.5.2
1.5.3
Methionine synthase, the canonical base-off Cbl binding protein
Methylmalonyl-CoA mutase: a large-scale conformational change
upon substrate binding
Diol and glycerol dehydratases: Base-on enzymes with a catalytic
K+ ion in their active sites
1.5.4
Glutamate mutase: a theory of radical propagation
1.5.5
1.5.6
Class II ribonucleotide reductase: a ten-stranded o# barrel
Lysine 5,6-aminomutase
1.6 AdoMet radical enzymes...................................................................................... 32
1.6.1 Lysine 2,3-aminomutase: reversible cleavage of AdoMet
1.6.2
Pyruvate formate lyase activase: a model for AdoMet binding
1.6.3
Class III ribonucleotide reductase activase: reductive cleavage of
AdoMet
1.6.4 Biotin and lipoyl synthases: sulfur insertion enzymes
1.6.5 Coproporphyrinogen III oxidase: a new fold for an Ado. enzyme
1.7 Summary ............................................................
1.8 Figures ............................................................
1.9 References ............................................................
38
39
54
Chapter 2: Preliminary crystallographic characterization of biotin synthase
Abbreviations...................................................................................................... 66
2.1 Protein purification, Fe-S cluster reconstitution, and addition of substrates....... 67
2.2
2.3
2.4
2.5
Crystallization and preliminary analysis..............................................................
Tables ............................................................
Figures ............................................................
References ............................................................
-6-
67
76
81
87
Chapter 3: Crystal structure of biotin synthase
Abbreviations.......................................................................................................
3.1 Introduction...................................................................
89
................................... 90
3.2 Materials and methods.........................................................................................
3.3 Results ............................................................
93
95
3.3.1 Overall structure and the location of Fe-S clusters
3.3.2 A PLP binding site is not observed in the structure
3.3.3 The novel 2Fe-2S cluster of BioB
3.3.4 The 4Fe-4S cluster and AdoMet
3.3.5 The methionyl moiety
3.3.6 The dethiobiotin binding site
3.4 Discussion............................................................
101
3.4.1
Arginine as an Fe ligand
3.4.2
3.4.3
Structural insight into the reaction catalyzed by BioB
Questions arising from the structure of BioB
3.5 Tables ............................................................
3.6 Figures ............................................................
3.7 References ............................................................
105
109
122
Chapter 4: Crystal structure of lysine 5,6-aminomutase
Abbreviations.......................................................................................................
127
4.1 Introduction............................................................
..................................... 128
4.2 Materials and methods......................................................................................... 131
4.2.1 Protein preparation
4.2.2 Storage, handling, and crystallization of protein samples, and crystal
cryoprotection
4.2.3 Crystal properties, phasing, model building, and refinement
4.3 Results ............................................................
134
4.3.1 Overall structure
4.3.2 PLP-protein interactions
4.3.3 Cobalamin-protein interactions
4.3.4 The CxxCxxxC motif of lysine 5,6 aminomutase
4.4 Discussion............................................................
140
4.4.1 A hypothetical conformational change to bring AdoCbl into the
active site
4.4.2 Questions arising from the structure of 5,6-LAM
4.5 Tables ............................................................
4.6 Figures ............................................................
4.7 References ............................................................................................................
-7-
144
147
164
Chapter 5: Conclusions
Abbreviations.......................................................................................................
5.1 Structural comparison of BioB and coproporphyrinogen III oxidase..................
5.2 Structural comparison of BioB with diol dehydratase.........................................
5.3 A novel role for PLP in a 5'-deoxyadenosyl radical enzyme...............................
5.4 5'-deoxyadenosyl radical enzymes: barrel structures with reaction-specific
accommodations...................................................................................................
5.5 Figures ............................................................
5.6 References ............................................................
166
167
168
169
169
171
173
Appendices
1 Summary of 5,6-LAM crystallization experiments .............................................
174
2 Curriculum Vitae.........................
175
...................................
-8-
List of Tables
Table 2.1. Timeline of BioB crystallization..................................................................
76
Table 2.2. Hampton Research screens used to find BioB crystallization conditions....
Table 2.3. Additives used to optimize BioB crystallization conditions........................
Table 2.4. Data collected from a single crystal grown in condition BioB3.................
Table 2.5. Data collected from crystals grown in condition BioB5 and used to solve
the structure of BioB.......................
....................................
77
78
79
Table 3.1. Maximal BioB activity.................................................................................
105
80
Table 3.2. Data collection and refinement statistics..................................................... 106
Table 3.3. Metal-nitrogen (guanidine) distances from crystallographic structures of
metal-guanidine complexes........................................................................................... 107
Table 3.4. Fe-S cluster-ligand distances.......................................................................
107
Table 3.5. Electrostatic interactions and potential hydrogen bonds between BioB
and AdoMet...........................................................
108
Table 3.6. Potential hydrogen bonds between BioB and DTB.................................... 108
Table 4.1. 5,6-LAM data collection and refinements statistics..................................... 144
Table 4.2. Potential PLP-5,6-LAM hydrogen bonding and electrostatic interactions.. 145
Table 4.3. Potential AdoCbl-5,6-LAM hydrogen bonding interactions....................... 146
-9-
List of Figures
Figure 1.1. Molecular structures of AdoCbl, Ado., and AdoMet................................ 39
Figure 1.2. Simplified catalytic cycle of AdoCbl-dependent mutases.......................... 39
Figure 1.3. Selected AdoMet-dependent enzymatic reactions .................................... 40
Figure 1.4. The biotin biosynthetic pathway.................................................................
41
Figure 1.5. The AdoCbl-dependent enzymatic reactions.............................................. 42
Figure 1.6. Several moderately or well -characterized AdoMet radical enzyme
reactions........................................................................................................................
43
Figure 1.7. The AdoMet cycle.............................................................
Figure 1.8. Crystal structure of the MeCbl-binding domain of methionine synthase...
Figure 1.9. Crystal structure of methylmalonyl-CoA mutase .......................................
Figure 1.10. Crystal structure of diol dehydratase........................................................
Figure 1.11. Crystal structure of glutamate mutase......................................................
Figure 1.12. Crystal structure of class II ribonucleotide reductase...............................
Figure 1.13. Crystal structure of the substrate-free AdoCbl-PLP-5,6-LAM complex..
Figure 1.14. Radical generation by the 4Fe-4S cluster and AdoMet in the AdoMet
radical enzymes.............................................................................................................
Figure 1.15. Mechanism proposed by Frey and coworkers for the 2,3-LAM
44
45
46
47
48
48
49
reaction..........................................................................................................................
51
50
Figure 1.16. Crystal structure of BioB .......................................................................... 52
Figure 1.17. Crystal structure of HemN ........................................................................ 53
Figure 2.1. SDS-PAGE analysis of BioB crystals grown in condition BioB2.............
Figure 2.2. Lattice contacts in the final structure of BioB ............................................
Figure 2.3. BioB crystals...............................................................................................
Figure 2.4. First diffraction from a BioB crystal..........................................................
Figure 2.5. 5.5 Adata from crystals grown in condition BioB3...................................
Figure 2.6. Location of putative Fe sites in BioB.........................................................
81
82
83
84
85
86
Figure 3.1. Overall reaction of BioB .............................................................
109
Figure 3.2. Mechanisms proposed for the BioB reaction..............................................
Figure 3.3. Stereo figure showing the electron density of the Fe-S clusters of BioB...
Figure 3.4. Solvent flattened, two-fold NCS-averaged 3.7 A resolution experimental
electron density map of BioB .............................................................
Figure 3.5. Ramachandran plot of the BioB monomer.................................................
110
111
Figure 3.6. Overall structure of BioB .............................................................
Figure 3.7. The BioB monomer....................................................................................
113
115
111
112
Figure 3.8. Space-filling model of BioB...................................................................... 116
Figure 3.9. Lattice packing in BioB crystals.................................................................
117
Figure 3.10. Stereo alpha carbon trace of a single BioB subunit, showing the
locations of all Lys residues in the structure.................................................................
117
118
Figure 3.11. The BioB active site.............................................................
Figure 3.12. 2Fe-2S and 4Fe-4S clusters with incomplete cysteinyl ligation............... 119
Figure 3.14. Conservation of residues across the AdoMet radical enzymes................
120
-10-
Figure 3.15. The carboxylate tail of DTB interacting with residues T292 and T293... 121
Figure 4.1. Aminomutases in the bacterial lysine fermentation pathway..................... 147
Figure 4.2. Proposed mechanism of 5,6-LAM .............................................................
148
Figure 4.3. 1,2-imino shift under radical-generating conditions and rearrangement
of the aziridylcarbinyl radical....................................................................................... 149
Figure 4.4. Various crystal forms of 5,6-LAM .............................................................
150
Figure 4.5. Solvent-flattened 3.0 Aresolution experimental electron density of 5,6LAM .............................................................
Figure 4.6. Ramachandran plot of 5,6-LAM .............................................................
152
153
Figure 4.7. Overall structure of 5,6-LAM and location of cofactors............................
Figure 4.8. Topology diagram of the 5,6-LAM ca3heterodimer..................................
Figure 4.9. Lattice packing interactions for 5,6-LAM .................................................
Figure 4.10. The five established PLP-binding enzyme families, plus 5,6-LAM ........
Figure 4.11. Superposition of the TIM barrels of alanine racemase and 5,6-LAM......
154
155
156
157
159
Figure 4.12. PLP in the putative active site of 5,6-LAM ..............................................
160
Figure 4.13. The AdoCbl binding site of 5,6-LAM ..................................................... 161
Figure 4.14. A stereo view of the local structure of the CxxCxxxC sequence .............
161
Figure 4.15. Hypothetical conformational change of 5,6-LAM and observed
conformational change of MCM .............................................................
162
Figure 4.16. A cleft leading to the putative active site of 5,6-LAM............................. 163
Figure 5.1. Superposition of BioB and coproporphyrinogen III oxidase ......................
Figure 5.2. Superposition of BioB and diol dehydratase ..............................................
- 11 -
171
172
Chapter
1
Introduction to AdoCbl and AdoMet -dependent radical
enzyme structures
- 12-
Abbreviations:
2,3-LAM
Lysine 2,3-aminomutase
5,6-LAM
Lysine 5,6-aminomutase
Ado
Ado.
AdoCbl
AdoH
AdoMet
APCbl
BioB
Cbl
CNCbl
CoA
Cob(II)
DDH
DTB
EAL
ENDOR
EPR
ESEEM
EXAFS
GDH
GM
HemN
ICM
LipA
MCM
MGM
MS
OAM
PFL
RNR
5'-deoxyadenosyl group
5'-deoxyadenosyl radical
Adenosylcobalamin, Coenzyme B 12
5'-deoxyadenosine
S-adenosylmethionine
Adeninylpentylcobalamin
Biotin synthase
Cobalamin
Cyanocobalamin, Vitamin B1 2
Coenzyme A
Cob(II)alamin
Diol dehydratase
Dethiobiotin
Ethanolamine ammonia lyase
Electron nuclear double resonance
Electron paramagnetic resonance
Electron spin-echo envelope modulation
Extended X-ray absorption fine structure
Glycerol dehydratase
Glutamate mutase
Coproporphyrinogen III oxidase
Isobutyryl-coenzyme A mutase
Lipoyl synthase
Methylmalonyl-coenzyme A mutase
Methyleneglutarate mutase
Methionine synthase
Omithine aminomutase
Pyruvate formate lyase
Ribonucleotide reductase
SAH
S-adenosylhomocysteine
SAM
S-adenosylmethionine
- 13-
Certain enzymatic reactions require the activation of saturated carbons in order to effect
difficult reactions. One general mode of accomplishing this daunting task is to recruit
metallocofactors, which supply reactivity beyond that of the set of naturally occurring
amino acids and organic cofactors. Carbon-based radical intermediates are sometimes
involved in such metal-based activation processes.
Two particular cofactors are
frequently recruited to abstract a hydrogen atom from an organic substrate, generating
free radical intermediates. The first of these, adenosylcobalamin (AdoCbl, or coenzyme
B12 , Fig. 1.la), is the coenzyme form of the cobalt-containing vitamin B1 2 , one of the
oldest known organometallic compounds. A second common radical-generating cofactor
is S-adenosylmethionine
(AdoMet or SAM, Fig. 1.1c) in complex with a 4Fe-4S cluster.
The immediately obvious feature that is common to both cofactors is the adenosyl group.
1.1 AdoCbl-dependent enzymes
The clinical importance of vitamin B 12 is rooted in the history of pernicious anemia, an
autoimmune disease of the gastrointestinal system that results in poor nutrient absorption.
If left untreated, the condition invariably results in death. The first reports of the disease
date back to the
1 8 th
and
1 9 th
centuries.
In 1926, Minot & Murphy demonstrated that
pernicious anemia could be cured by feeding patients large amounts of liver2 , a discovery
that would later win them the Nobel Prize in Medicine and Physiology. The groups of
Folkers3 and Lester-Smith4 eventually isolated cyanocobalamin (CNCbl or vitamin B12 ),
a derivative of the "anti pernicious anemia factor" present in liver. Other serious human
diseases arising from vitamin B 12 deficiency were later discovered, including
methylmalonic aciduria and homocysteinuria. Both of these are genetic disease caused
-14-
by mutation in either the MutAB genes5 ,
which encode
AdoCbl-dependent
methylmalonyl-CoA mutase (MCM), or in one of the MmmA-H genes, which encode
proteins associated with Cbl uptake and metabolism6 .
In 1958, Barker and coworkers showed that crude cell extracts of Clostridium
tetanomorphum required a coenzyme form of vitamin B 12 to convert glutamate to 3methylaspartate. This work led to the discovery and purification of AdoCbl7. AdoCbl
dependence in a purified enzyme system (MCM) was first reported by Stadtman and
coworkers8' 9 two years after Barker's discovery of the coenzyme. Eggerer et al proposed
that AdoCbl-dependent MCM catalyzes the formation of paramagnetic species on the
reaction pathway9 . Their hypothetical mechanism fell short of recognizing C-Co bond
cleavage, (Fig. 1.2) but made an intellectual leap forward in asserting that the
rearrangement of methylmalonyl-CoA to succinyl-CoA involved radical intermediates.
Thirty years after Minot and Murphy's successful treatment of pernicious anemia,
Hodgkin and coworkers solved the X-ray crystal structure of CNCbl l° , for which she was
awarded the 1964 Nobel Prize in Chemistry. The structure revealed a single octahedral
Co(III) center, coordinated equatorially by the four pyrrole nitrogens of a corrin
macrocycle.
The lower axial ligand is a nitrogen atom of the cofactor's
dimethylbenzimidazole moiety, and the upper axial ligand is cyanide, an artifact of
purification.
In 1961, Hodgkin solved the structure of the coenzyme form of B 12 ll ,
illuminating the structural basis for the reactivity of AdoCbl. At the heart of AdoCbl's
reactivity is its alkyl cobalt bond: the 5'-deoxyadenosyl group (Ado) is covalently bound
- 15-
to Co(III) in the upper axial position. AdoCbl is one of the largest known biological
cofactors; it has the chemical formula C72H100 N1 801 7 PCO,and contains 18 chiral centers.
In 1966, R.H. Abeles and P.A. Frey first showed that diol dehydratase (DDH) catalyzes
the transfer of tritium atoms between [1-3 H]propanediol and the 5' carbon of AdoCbl' 2 .
Furthermore,
the resultant
[5'-3 H]AdoCbl
was able to transfer
its tritium
into
propionaldehyde, the product of the diol dehydratase reaction. In the early 1970's, other
key investigators of AdoCbl enzymes, including Blakley and Babior, observed the
kinetically competent formation of a paramagnetic species in the reactions of class II
RNR 13
and EAL ' 41
5.
These and other pioneering
experiments,
over years of
investigation, have led to the hypothesis that the C-Co bond of AdoCbl is homolytically
cleaved, resulting in the transient intermediates Cob(II)alamin [Cob(II)] and 5'deoxyadenosyl radical (Ado., Fig. l.lb and Fig. 1.2), which is thought to be responsible
for the H atom abstractions that initiate AdoCbl-dependent reactions. Several excellent
reviews covering the reactivity of AdoCbl enzymes are available16
23 .
1.2. AdoMet radical enzymes
Twenty three years after demonstrating the 3H atom transfers in DDH, Frey published
analogous experiments on AdoMet-dependent lysine 2,3-aminomutase24 (2,3-LAM),
supporting the hypothesis that Ado. is the agent that is responsible for H atom
abstraction in the enzymatic reaction. Significantly, Barker's work on 2,3-LAM first
demonstrated that the enzyme is pyridoxal 5'-phosphate-(PLP) and Fe -dependent2 5.
- 16-
Compared to AdoCbl, AdoMet is a relatively simple molecule (Fig. 1.1). The C-Co bond
of AdoCbl (-30 kcal/mol) is replaced by a much stronger (-60 kcal/mol) bond between
the 5' carbon of Ado and the sulfur atom of the methionyl group. AdoMet is involved in
a number of unusual reactions (Fig. 1.3). Two such reactions appear in the biotin
(vitamin B8) biosynthetic pathway (Fig. 1.4). In the first of these reactions, AdoMet acts
as an amino
group
donor
in
the PLP-dependent
transformation
of
7-keto-8-
aminoperlargonic acid to 7,8-diaminopelargonic acid2 6 . The second unusual reaction is
that of biotin synthase (BioB). BioB is an iron-dependent protein that requires AdoMet
to form biotin from dethiobiotin 27
29
(DTB).
Marquet and coworkers have shown that
BioB transfers deuterium from deuterated DTB into 5'-deoxyadenosine (AdoH)3 0 . These
deuterium transfer experiments support the hypothesis that AdoMet abstracts substrate
hydrogen atoms in BioB, as it does in 2,3-LAM.
1.3. Why two radical cofactors?
Since the early work on AdoCbl enzymes, a total of 10 AdoCbl- and 6 AdoMet-
dependent radical enzymes have been purified and reasonably well characterized (Figs.
1.5 and 1.6). Recently, Sofia and coworkers proposed that over 600 unique sequences of
AdoMet radical enzymes exist31 . Several fascinating questions arise from this work:
Why has nature evolved two analogous cofactors to abstract H atoms from metabolic
intermediates? What is the evolutionary relationship between the two cofactors and the
enzymes that require them?
What are the evolutionary determinants of cofactor
preference for a given enzyme? It is tempting, upon initial inspection, to speculate that
- 17-
the AdoMet radical enzymes are the molecular ancestors of the AdoCbl radical enzymes.
Proponents of this theory may point to a few salient facts:
1.
All of the known AdoMet radical enzymes require a 4Fe-4S cluster and, with one
known exception, anoxic conditions for activity. Fe-S clusters are widely regarded
as ancient and ubiquitous cofactors, capable of carrying out catalysis in the
32
ancient, anoxic atmosphere
2.
34 .
AdoMet biosynthesis (Fig. 1.7) requires far fewer enzymes than does AdoCbl
biosynthesis (this argument does not take into account the enzymes required for
Fe-S cluster biosynthesis or cluster reduction).
3.
Several equivalents of AdoMet are required as methylating agents in the
biosynthesis of AdoCbl3 5.
4.
Sequence analysis on probable AdoMet radical enzymes by Sofia et al reveals that
some of these enzymes contain part of the "base-off' AdoCbl-binding sequence
motif (DxHxxG...Sxl...GG)
that is found in all of the AdoMet-dependent
mutases.
5.
Several AdoMet radical enzymes catalyze steps in central pathways of
metabolism, including DNA biosynthesis and repair, acetyl-CoA biosynthesis, and
coenzyme biosynthesis.
These and other reasons have led some to classify AdoMet radical enzymes as the
"molecular fossils" of AdoCbl radical enzymes. In particular, this speculation has been
applied with respect to the ribonucleotide reductases36 3 7. Others believe that AdoCbl
- 18-
may be the ancestral radical cofactor, and that the various arguments of the AdoMet
ancestor proponents are weak. This opinion is bolstered by several lines of reasoning:
1.
Corrinoid-like molecules are thought to date back to the primordial soup3 8 .
2.
Many (but not all) organisms require methylcobalamin to synthesize AdoMet via
the cobalamin-dependent methionine biosynthetic pathway3 9.
3.
The presence of the "base-off' AdoCbl binding motif in AdoMet radical enzyme
sequences implies no directionality in evolution.
4.
Several AdoCbl-dependent enzymes catalyze steps in central pathways of
metabolism, including DNA biosynthesis, such as the class II ribonucleotide
reductase (RNR).
5.
In comparison to the AdoMet radical enzyme activases, where an entire protein,
AdoMet, and a 4Fe-4S cluster are required to generate a catalytic radical on a
separate enzyme, one can argue that AdoCbl is the simpler radical-generating
machinery.
Some AdoCbl-dependent
enzymes, such as the class II RNR, lysine 5,6-aminomutase
(5,6-LAM), and diol dehydratase (DDH) have AdoMet-dependent counterparts that
catalyze similar reactions.
Perhaps these enzymes provide the best opportunity for
understanding how the two enzyme superfamilies may be related.
-19-
1.4. The diversity of 5'-deoxyadenosyl radical enzymes
Many AdoCbl radical enzyme share a set of common steps in their reaction mechanisms,
shown in general terms in Fig. 1.2. Briefly, Ado. abstracts a non-activated hydrogen
atom from the substrate. The substrate then rearranges via a vicinal group migration to
the original radical center, and then re-abstracts a hydrogen atom from AdoH, resulting in
the reformation of AdoCbl. The exception to this general mechanistic scheme is the class
II RNR reaction, in which the radical is propagated
to a Cys residue, generating a
catalytic thiyl radical'3' 40' 41. A key question in the field is how the AdoCbl enzymes, in
the presence of either substrate or effectors, accelerate C-Co bond homoloysis by a factor
of-101 2 over the non-enzymatic cleavage174246.
In contrast, AdoMet radical enzymes catalyze more diverse reactions (Fig. 1.6), and have
substrates that vary in size from small molecules to proteins to DNA. The reaction
mechanisms of AdoMet-dependent radical enzymes cannot be generalized beyond the
abstraction of an H atom by Ado.. While some AdoMet radical enzymes go through a
reaction sequence similar to those of AdoCbl enzymes, others utilize AdoMet as a
substrate rather than a cofactor, resulting in the production of one or more equivalents of
methionine and AdoH per substrate consumed.
Why are AdoMet radical enzymes so much more versatile than AdoCbl radical enzymes?
What are the implications of this diversity on the theory that AdoMet is the ancestral
radical cofactor? Given the current state of the field, the answers to these and other
- 20 -
questions about the relationship between AdoCbl and AdoMet radical enzymes remain
unknown. Some of the answers may lie in the structural details of the two superfamilies.
1.5. AdoCbl and AdoMet -dependent radical enzymes: structural information
This section briefly reviews some of the more well-characterized AdoCbl and AdoMetradical enzymes, with emphasis on structural information. The study of several of these
enzymes has benefited from the availability of X-ray or NMR structures. A unifying
theme in the structures of these enzymes is the dominance of a/odbarrel folds4 7, in which
a central f-sheet is surrounded by a helices, and the prevalence of the TIM barrel fold in
particular.
The TIM barrel is an ancient, ubiquitous, and adaptable protein fold4 8 4 9. TIM barrel
proteins carry out an astonishing number of diverse metabolic functions. Twenty-six
protein superfamilies use the TIM barrel architecture to bind their substrates and to place
their catalytic elements in their active sites47.
There is no apparent theme to the
enzymatic reactions carried out by TIM barrels, and at least one TIM barrel has no
associated enzymatic activity5° .
With the exception of class II RNR, all AdoCbl-
dependent radical enzymes of known structure have the same protein motifs, consisting
of a TIM barrel substrate-binding domain and an AdoCbl binding domain that docks on
top of the TIM barrel. The work described in this thesis reveals that BioB, an AdoMet
dependent radical enzyme, is a TIM barrel. Furthermore, the structure of 5,6-LAM, an
enzyme which bridges the AdoCbl and AdoMet radical superfamilies, adheres to the
general scaffold of AdoCbl-dependent radical enzymes.
-21 -
One interesting conclusion that is illuminated by structural and biophysical studies of
cobalamin- (Cbl) binding proteins is that they can be grouped based on their mode of
binding to the protein.
The group I ("base-off") proteins use a histidine residue to
displace the intrinsic DMB moiety of the cofactor as the lower axial ligand to cobalt.
This group includes methylcobalamin-dependent methionine synthase (MS) 51 , MCM 52 ,
GM53 , 5,6-LAM54 (this work), and, presumably, isobutyryl-CoA mutase (ICM)55,
methyleneglutarte mutase (MGM)56 , and omithine aminomutase (OAM)57 . For the group
I enzymes that utilize AdoCbl, catalysis is initiated by abstraction of a H atom from a
non-activated substrate carbon having only hydrocarbon substituents.
The group II
("base-on") proteins, which include DDH58 , EAL59, class II RNR60 , the periplasmic
CNCbl-binding
protein BtuF 61, and the ATP:corrinoid adenosyltransferase
62 ,
with the DMB moiety in place as the lower axial ligand. The AdoCbl-dependent
bind the
group II
enzymes initiate catalysis by abstracting H from a heteroatom (class II RNR) or from a
carbon with heteroatom substituents (EAL, DDH, GDH).
1.5.1.Methioninesynthase,the canonicalbase-off Cblbindingprotein
In 1994, Drennan et al solved the crystal structure of the Cbl-binding domain of MS51
(PDB code 1BMT), providing the first structure of protein-bound Cbl and the canonical
model of a base-off Cbl-binding enzyme (Fig. 1.8a). The structure of the Cbl-binding
subunit is a Rossmann-like fold, which is the norm for base-off Cbl binding. One of the
most important results of this study was the establishment of the structural role of the
base-off Cbl binding sequence DxHxxG...Sxl...GG (Fig. 1.8b). This sequence is found
in all base-off Cbl-binding enzymes. The DxH portion is involved in Co ligation, where
- 22 -
the Asp sidechain hydrogen bonds to the imidazole ring of the His ligand.
The
downstream Gly residue leaves a cavity in which the phosphate of the Cbl nucleotide can
bind. The Ser residue of the Sxl portion hydrogen bonds to imidazole ring of DMB, and
the two Gly residues at the end of the sequence pack against the DMB moiety.
1.5.2.Methylmalonyl-CoAmutase: a large-scaleconformationalchange upon substrate
binding
MCM catalyzes the interconversion of methylmalonyl-CoA and succinyl-CoA (Fig.
1.5a). From a structural perspective, MCM is one of the most thoroughly characterized
AdoCbl enzymes. Two years after the publication of the structure of the Cbl-binding
portion of MS, Mancia et al published the structure of MCM in complex with a substrate
fragment and with AdoCbl52 (PDB code 1REQ; Fig. 1.9a). The Ado moiety was not
modeled in this structure due to its poor electron density. As in MS, a His residue
coordinates the Co of Cbl. Remarkably, homolysis of the C-Co bond of the enzymebound cofactor is accelerated by approximately a trillion-fold42 4 5. Unlike MS, the His NCo distance in MCM is somewhat long (2.5 A) compared with the DMB N-Co distance
in the crystal structure of free cofactor6 3 (2.237 A), which lead the authors to conclude
that trans effects646 5 in the protein-bound cofactor were responsible for promoting C-Co
bond homolysis when substrate is bound. Structural comparison reveals that the lower
axial His ligand of MCM and the Asp sidechain that hydrogen bonds to it superimpose
with the His/Asp couple in MS, and that relative to MS, the cofactor has elevated B
factors and is slightly farther away from the Rossmann domain. Thus, it is not protein
differences that give rise to variations in N-Co bond length, but rather cofactor disorder
- 23 -
that appears to be responsible. One source of cofactor disorder may be X-ray irradiation,
which is also responsible for the loss of the Ado ligand6 6 in this crystal structure.
Regardless of the source of cofactor disorder, the long N-Co bond length reported for
MCM is not likely to be catalytically relevant, since extended X-ray absorption fine
structure (EXAFS) analysis67 does not support the trans effects hypothesis in MCM.
The -23 A-long substrate fragment desulfo-CoA (coenzyme A without the terminal thiol
group) is bound along the entire length of the TIM barrel axis, with the end that would
contain the thiol located -11 A from the cobalt of the cofactor, near the C-terminal end of
the barrel. This suggested that the true substrate binds with its methylmalonyl group
positioned near the Ado moiety of AdoCbl.
Indeed, when a combination of true
substrate/product was observed in the structure of the AdoCbl-MCM complex 68 (PDB
code 4REQ), proximity of the substrate/product to AdoH was confirmed, as well as the
identification of several key residues that bind the true substrate. H244 interacts with the
carbonyl of the thioester, consistent with the thought that partial protonation of the
carbonyl oxygen of the migrating group may facilitate the reaction 69 71 (see Fig. 1.5a).
Mutation of this residue results in a -103-fold reduction in kcat.
The crystal structure of the substrate-free MCM-AdoCbl complex68 (PDB code 3REQ;
Fig. 1.9b) revealed a drastically different conformation of the enzyme. Remarkably, the
TIM barrel is distended to the point of rupture. The carbonyl-amide backbone hydrogen
bonds of the internal beta sheet cylinder are interrupted by a gap in the side of the barrel,
leaving enough room for substrate to enter the TIM barrel. The overall B factor of 80 A2
-24-
suggests that this structure is conformationally dynamic. Unlike crystals of the desulfoCoA MCM complex, the spectrum of substrate-free crystals was consistent with the
presence of Co(III), which is expected of an intact, six-coordinate AdoCbl species.
Features of electron density consistent with an Ado group still bonded to Co allowed
Mancia and Evans to model the intact cofactor in this structure. Contrary to the previous
hypothesis concerning trans effects in the presence of substrate, there was no evidence
for a shorter Co-N bond length in the substrate-free structure. As in the desulfo-CoA
MCM complex, the B factors of the cofactor are high. Finally, Mancia and Evans
published structures of two inhibitor-MCM complexes (PDB codes 6REQ, 7REQ)72 . In
the CoA MCM complex, the CoA is disordered and the enzyme adopts a conformation
like that of the open-barrel, substrate-free MCM-AdoCbl structure.
One of the most interesting results of the structural work on MCM is the observation of a
conformational change of Y89, a conserved residue of the Cbl-binding site. In the
substrate-free structure, Y89 is located above the intact cofactor, with its aromatic ring
roughly parallel to the adenine of Ado. In the substrate-bound structure68 (PDB code
4REQ), the sidechain of Y89 occupies the former Ado site and hydrogen bonds to the
substrate.
This destroys the Ado binding site, and is thought to be a major factor
contributing to the rate acceleration of C-Co bond homolysis in MCM. The significance
of the hydrogen bond between the phenolic oxygen of Y89 and the substrate is reflected
in kinetic studies of the Y89F mutant, which shows a 1000-fold reduction in kat73 .
Moreover, formation of Cob(II) in either the pre-steady state or in the steady state is not
observed in the Y89F mutant73 , as it is with the wild-type protein.
- 25 -
The structural work on MCM, combined with biochemical studies, gives us a more
complete picture of how an AdoCbl enzyme accelerates C-Co bond homolysis in the
presence of substrate. The substrate free structure adopts a conformation which allows
the substrate to enter the active site, causing closure of the TIM barrel. A change in the
conformation of Y89, caused by substrate binding, destroys the Ado binding site and
seems to play a key role in C-Co homolysis.
1.5.3. Diol and glycerol dehydratases:Base-on enzymes with a catalyticK+ ion in their
active sites
Diol dehydratase (DDH) was first purified and characterized by Lee and Abeles, who
found a monovalent cation requirement for activity7 4 . DDH and glycerol dehydratase
(GDH) are similar but distinct enzymes that catalyze a 1,2 hydroxyl group shift followed
by the elimination of water to give an aldehyde product75 (Fig. 1.5e, f). The structure of
DDH in complex with CNCbl and (S)-1,2-propanediol76 (PDB code 1DIO), published in
1999 by Yasuoka and coworkers, highlighted several fascinating aspects of this enzyme.
Most strikingly, Cbl is bound in the base-on conformation (Fig. 1.10Oa). The protein is a
dimer of heterotrimers, (ay) 2, which binds two equivalents of AdoCbl. Cbl is bound by
the
subunit, with the upper face of the corrin ring projected into the substrate-binding
TIM barrel (a subunit). The y subunit is proposed to stabilize the heterotrimer. Like
MCM, a long N-Co bond (2.5 A) was observed for the structure of the DDH-CNCbl
complex.
The upper axial CN ligand was not modeled due to poor electron density,
presumably caused by radiolytic damage.
- 26 -
The base-on configuration of DDH-bound Cbl requires a Cbl binding subunit that is
different from the Rossmann-like domain that invariably binds Cbl in MS and the
mutases (Fig. 1.10b). The fi subunit of DDH consists of a central -sheet with peripheral
c-helices.
The cofactor binds on the face of the
-sheet, flanked by two helices.
The
orientation of the cofactor with respect to the fl-sheet does not resemble the arrangement
of AdoCbl in the Rossmann domains of base-off enzymes.
The crystallographic data indicate the presence of a K+ ion in the active site, located deep
in the core of the TIM barrel (Fig. 1.10b). The K+ ion is coordinated by the hydroxyl
groups of the substrate, implying that it participates directly in catalysis76 , perhaps by
lowering the energy barrier to a hypothetical oxycation radical transition state (Fig.
1.10c). The observation of a K+ ion in the active site is consistent with Abeles' initial
finding of a monovalent cation requirement for DDH74 . Interestingly, in the substratefree DDH-CNCbl complex, two water molecules coordinate the K+ ion, rather than the
hydroxyl groups of 1,2-propanediol. A similar observation was made by Liao et al in
their structural studies of substrate-bound vs. substrate-free GDH77 .
Structural comparison of the DDH-adeninylpentylcobalamin (APCbl; an AdoCbl analog)
substrate complex58 (PDB code EEX) with the substrate-free DDH-CNCbl complex7 8
(PDB code 1IWB) reveals that the f subunit and the corrin macrocycle move away from
the a subunit when substrate binds. Assuming that the Ado moiety is held in the same
position as in the substrate-free structure, movement of the corrin macrocycle would
lengthen the C-Co bond and cause its cleavage. This idea has been referred to as an
- 27 -
adenine-attracting effect; the affinity for the adenine moiety is so strong that when the
Cbl-binding domain moves away as substrates bind, the Ado group is held in place and
the C-Co bond cleaves.
For both MCM and DDH, conformational changes upon substrate binding suggest
mechanisms by which the proteins can enhance the rate of C-Co bond homolysis. For
MCM, substrate binding causes the dramatic closure of the (a/gJ)8 barrel and the
movement of Tyr 89 into the Ado-binding site, displacing Ado from the upper axial
position. In DDH, the enzyme binds both ends of the AdoCbl cofactor, and substrate
binding causes the Ado moiety and the Cbl portion to pull away from each other.
Interestingly, prior to crystallographic analyses of Cbl-dependent enzymes, mechanisms
to explain the increased rates of C-Co bond homolysis of the enzyme-bound cofactor
focused on fine-tuning of the nucleophilicity of the lower axial ligand to Co. Structural
data has shifted the focus to "brute-force" mechanisms for C-Co bond homolysis rate
enhancement.
1.5.4. Glutamatemutase:a theory of radicalpropagation
Glutamate mutase (GM) catalyzes the interconversion of (S)-glutamate and (2S,3S)-3methylaspartate, facilitating the fermentation of glutamate to acetate, CO2, and NH379
(Fig. 1.5c). The enzyme is a heterotetramer,
2a2 , where
is a large subunit and a is a
small subunit. Kratky and coworkers have determined structures of GM in complex with
AdoCbl and substrate8 0 (PDB code 119C;Fig. 1.11), CNCbl and a substrate mimic (PDB
code 1CCW) 53 , and MeCbl and a substrate mimic5 3 (PDB code 1CB7). GM binds two
- 28 -
AdoCbl molecules, one in each Rossmann-like domain (a subunit). Like all of the
AdoCbl-dependent mutases, GM binds AdoCbl in the base-off conformation. EXAFS
studies of GM do not indicate an unusually long N-Co bond66 , which is confirmed by the
crystallographic bond length of 2.2 A. The top face of the corrin macrocycle is presented
at the C-terminal end of the substrate-binding TIM barrel ( subunit). Arguably, the most
important result from the crystallographic analysis of GM is the observation of two
conformations of the Ado moiety of AdoCbl. The 1.9 A resolution structure suggests a
mixed conformation for Ado, with one population in the C2'-endo conformation, and the
other in the C3'-endo conformation. The interconversion of these two conformations,
termed "ribose pseudorotation," is proposed to be the basis of radical propagation from
Ado. to the substrate. In the C2'-endo conformation, C5' is positioned -3.1 A directly
above the Co atom, whereas in the C3'-endo state, C5' is -4.5A removed from the Co and
positioned such that it can abstract an H atom from C4 of the substrate, as expected from
mechanistic studies 81.
In addition to the crystallographic structures, solution structures of the a subunit are
available82 8 5. The solution structures are generally similar to the X-ray structures, but
indicate that the loop containing the coordinating His residue, as well as the helix that
precedes this loop, are in a dynamic equilibrium between relatively structured and
unstructured states when AdoCbl is not bound. These studies suggest that the protein
locks onto AdoCbl and adopts a more rigid structure when the cofactor is bound.
- 29 -
1.5.5. ClassII ribonucleotidereductase:a ten-strandedao0barrel
Class II RNR carries out a unique AdoCbl-dependent reduction of nucleoside
triphosphates or diphosphates (Fig. 1.5h). The class II RNR from L. Leichmannii, for
which a structure has been determined6 0, acts on nucleoside triphosphates. A system
capable
of
multiple
turnovers
requires
dithiols
or
the
physiological
thioredoxin/thioredoxin reductase system38687. In addition to nucleoside triphosphate
reduction, L. leichmanii class II RNR catalyzes the exchange of 5' hydrogens of AdoCbl
with solvent, in the presence of effector but not substrate88 89. A protein based, cysteine
derived thiyl radical has been shown to be generated in a kinetically competent
fashion40 '4 ' 86,and several other cysteine residues were shown to play a role in delivering
reducing equivalents to the active site86 .
Formation of the thiyl radical is AdoCbl
dependent. Whether or not formation of the thiyl radical is concerted with C-Co bond
homolysis has been the subject of several investigations2 3 9
Electron paramagnetic resonance (EPR) studies demonstrated that class II RNR from L.
leichmannii binds AdoCbl in the base-on configuration9 1. This was confirmed by the
crystal structure of the class II RNR in complex with APCbl (PDB code 1LIL; Fig.
1.12)60. The structure of the class II RNR is largely dissimilar to the known structures of
other AdoCbl binding enzymes, including the base-on enzyme DDH. Based on the
location of Cys 408, which gives rise to the catalytic thiyl radical, the active site is
located in the middle of a 10-stranded ol barrel. Cbl is bound to an "AdoCbl binding
region" above the barrel.
A structure of the apoenzyme is also available60 , and
comparison with the APCbl-bound structure reveals a clamping-down of the AdoCbl
- 30 -
binding region over the barrel when Cbl is bound. The solvent accessibility of the active
site Cys residue in the APCbl-bound structure suggests that the AdoCbl binding region is
not fully closed over the barrel, resulting in a -10
A separation between
the active site
Cys in the barrel and the Co atom of Cbl. EPR spectroscopy was used to show that the
Co-Cys distance is 5.5-7.5 A in the catalytically active enzyme-coenzyme complex41 .
The openness of the APCbl-RNR structure was attributed to lack of bound effector.
92
Effector binding promotes C-Co bond homolysis
93 ,
perhaps by causing a more
complete movement of the AdoCbl binding region, which would also bring the cofactor
and active site Cys residue into proximity. Since the adeninylpentyl moiety of APCbl
was not interpretable, the details of how clamping-down of the AdoCbl binding region
would effect C-Co bond homolysis are unknown at this time.
A striking feature of the class II RNR structure is that the global fold of the enzyme is an
(a13)
10 barrel, unlike any other AdoCbl enzyme of known structure. This structure is
similar to those of the class i94 and III95 RNRs and suggests that, as a class of enzymes,
the RNRs are more closely related to one another than the class II enzyme is to other
AdoCbl enzymes.
1.5.6. Lysine 5,6-aminomutase
5,6-LAM is an AdoCbl96 '99 and PLP dependent l '°° l° enzyme in the bacterial lysine
fermentation pathway0 2 . This enzyme catalyzes the interconversion of DL-lysine or of
fl-L-lysine to 2,5-diaminohexanoate
or to 3,5-diaminohexanoate,
respectively (Fig. 1.5i).
5,6-LAM is similar to 2,3-LAM in terms of its substrates and reaction, but 2,3-LAM is an
-31 -
AdoMet radical enzyme.
characterized
recombinant
Frey and coworkers have overexpressed, purified and
5,6-LAM
from
C. sticklandii54 ,103 .
Interestingly,
computational analyses suggest a role for PLP in stabilizing radical intermediates on the
reaction pathway'03 , though Danen and West demonstrated that the 1,2 amino shift occurs
in the absence of any aromatic substituent in the simplest model, the aziridylcarbinyl
radical' 0 4.
The crystal structure of 5,6-LAM (Fig. 1.13), a major contribution to this
thesis, was determined in collaboration with the Frey laboratory and suggests a novel role
for PLP. The structure will be discussed in further detail in chapter 4.
1.6. AdoMet radical enzymes
In this section, the well or moderately-well -characterized AdoMet radical enzymes are
introduced.
Only two X-ray structures for AdoMet radical enzymes are published,
though spectroscopic studies have provided structural insight into the coordination
environment of the 4Fe-4S clusters in these enzymes.
1.6.1. Lysine 2,3-aminomutase: reversible cleavage of AdoMet
2,3-LAM, another enzyme of the bacterial lysine fermentation pathway02 , was purified
by Barker and coworkers from Clostridium subterminale SB425 . The enzyme catalyzes
the interconversion of L-lysine and L-fl-lysine (Fig. 1.6c). 2,3-LAM is a homohexamer,
with each monomer containing 1 molecule of PLP0 5 , 1 molecule of AdoMet24
an oxygen labile 4Fe-4S cluster"l l° '.
'106 '10 9,
and
Like all known AdoMet radical enzymes, 2,3-
LAM has a conserved CxxxCxxC motif.
Frey and coworkers have made major
contributions to our understanding of 2,3-LAM, which is one of the best- characterized
- 32 -
AdoMet radical enzyme. As mentioned earlier in this chapter, 3H transfer from substrate
to AdoMet was demonstrated by Moss and Frey in 198724. Unlike many other AdoMet
radical enzymes, AdoMet cleavage is reversible in the 2,3-LAM reaction 10 9 , i.e. 2,3-LAM
utilizes AdoMet as a true cofactor rather than as a substrate. Cleavage of AdoMet to
Ado. and Met requires the 4Fe-4S cluster to be reduced to the +1 state, which itself
requires the presence of AdoMet T1 .
substrate08 .
AdoMet cleavage also requires the presence of
Spectroscopic investigations have led to a hypothetical mechanism for
reductive cleavage of AdoMet by the catalytic [4Fe-4S]1+ cluster (Fig. 1.14), in which the
AdoMet amino and carboxyl groups chelate a unique Fe of the cluster, and the AdoMet
sulfur is a ligand to the unique Fe after cleavage'0 6 '1 0 9.
Spectacular model chemistry1 12 and spectroscopicl3-116investigations with substrate and
with substrate analogs have allowed Frey and coworkers to propose a mechanism for 2,3LAM, in which PLP forms a benzylic radical at the imine carbon, facilitating a 1,2 imine
shift via an aziridylcarbinyl
radical (Fig. 1.15).
In the spectroscopic
studies, EPR,
electron spin-echo envelope modulation (ESEEM) spectroscopy, and rapid freeze quench
techniques were used to identify a kinetically competent radical intermediate in the
reaction of 2,3-LAM with substrate. The studies of 2,3-LAM have provided the basis for
many of the mechanistic proposals for 5,6-LAM, which will be discussed further in
chapter 4.
-33-
1.6.2. Pyruvateformate lyase activase: a modelfor AdoMet binding
PFL catalyzes the reversible conversion of pyruvate and CoA to acetyl CoA and formate,
an essential step in anaerobic glucose fermentation. PFL requires an activating enzyme,
first described by Knappe and coworkers' 7
9.
PFL activase contains a protein-bound
4Fe-4S cluster120' 1 21 , which is ligated by the CxxxCxxC motif, and is active in the 1+
oxidation state122 .
Cluster reduction is dependent on the flavodoxin/flavodoxin
reductase/NADPH system 23 . PFL activase abstracts an H atom from the accarbon of
PFL at Gly 734 (Fig. 1.6d), which is incorporated into AdoH, a product of the reaction24 .
PFL activase has been a model system for studying AdoMet-Fe-S cluster interactions in
AdoMet radical enzymes. Mossbauer studies by Huynh and Broderick revealed the
presence of a unique Fe in the 4Fe-4S cluster , and electron nuclear double resonance
(ENDOR) studies by Broderick and Hoffman demonstrated chelation of AdoMet to the
unique Fe of the cluster via the amino acid moiety 2 6 . In these ENDOR studies, the
AdoMet was labeled at the carboxyl group with 170 and ' 3 C, and at the amino group with
15N.
The model derived from these studies is shown in Fig. 1.14b. Although details of
the sulfonium interaction with the cluster remain to be established, this model appears to
represent the general mode of AdoMet binding in this superfamily, and has withstood the
test of additional spectroscopic studies on other AdoMet radical enzymes, and two crystal
structure determinations.
- 34-
1.6.3. ClassIII ribonucleotidereductaseactivase:reductivecleavageof AdoMet
Class III RNR from E. coli is responsible for deoxyribonucleotide biosynthesis under
anaerobic conditions. It was first described by Fontecave et al127 , and like PFL, it is a
glycyl radical enzyme. The reductase is activated by H atom abstraction (Fig. 1.6d)
under anaerobic conditions by an activase, generally called , which is dependent on one
equivalent of AdoMet'28 -'3 0 and contains a 4Fe-4S cluster ligated by a CxxxCxxC
motif 31
' 13 2.
EPR investigations on dithionite-reduced activase by Fontecave and
coworkers showed that the [4Fe-4S]l+ cluster reduces AdoMet to produce AdoH and
Met 129
' 30.
Flavodoxin/flavodoxin reductase/NADPH is thought to be the physiological
reducing system for the 4Fe-4S cluster and is competent for glycyl radical formation in
vitro3 3 . These experiments on the class III RNR were crucial in establishing the role of
the essential 4Fe-4S cluster in the AdoMet radical enzymes. Interestingly, Fontecave and
coworkers propose that reduction of the cluster by flavodoxin is thermodynamically
coupled to formation of AdoH and the glycyl radical on the target protein13 3 . A crystal
structure of the activase alone, and in complex with flavodoxin and the reductase, would
be invaluable for learning more about the mechanism by which this thermodynamic
coupling is achieved.
1.6.4. Biotin and lipoyl synthases: sulfur insertion enzymes
BioB catalyzes the final step in the biotin (vitamin B8 ) biosynthetic pathway (Fig. 1.4),
the conversion of dethiobiotin to biotin' 34 ' 13 5 (Fig. 1.6a).
BioB contains one 4Fe-4S
cluster and one 2Fe-2S cluster'36 . The 4Fe-4S cluster, which is responsible for radical
-35-
generation, is ligated by AdoMet and by the three cysteines of the CxxxCxxC motif,
3
which is common to all AdoMet radical enzymes
37 - 9.1
The 2Fe-2S cluster is ligated by
three Cys residues and one Arg residue13 7, and has been proposed to be the S donor in
the reaction'
37
'140-143,
but this is a subject of dispute144 ,145.
BioB is not known for its
catalytic prowess: the E. coli enzyme catalyzes only one turnover at a rate of minutes to
hours.
The only report of multiple turnovers is for BioB from Arabidopsis, which
catalyzes >2 turnovers per hour' 4 6 . The crystal structure of BioB' 3 7 (PDB code 1R30,
Fig. 1.16) is a major contribution to this thesis. The structure of BioB and the remarkable
reaction this enzyme performs will be discussed in further detail in chapter 3.
The lipoyl group is an essential cofactor in several important enzymes' 4 7 , including the
pyruvate dehydrogenase complex. Lipoyl synthase (LipA) catalyzes the conversion of
octanoylated proteins to lipoylated proteins 148' 15 (Fig. 1.6b). In many respects, LipA is
similar to BioB. Like BioB, LipA catalyzes the formation of C-S bonds at nonactivated
carbon positions and contains two Fe-S clusters per polypeptide (one of which is a 4Fe4S cluster that is, presumably, ligated by the protein's CxxxCxxC motif and by AdoMet).
Also like BioB (according to Marquet's report142),LipA consumes two equivalents of
AdoMet per product.
Despite the similarities of their reactions, some important
differences between LipA and BioB do exist. Whereas BioB produces one equivalent of
product per monomer, two equivalents of LipA are required to produce one lipoyl
product, and the sulfur insertion reaction proceeds with inversion of configuration4 9
in contrast to the retention of configuration
in BioB152.
As the biochemical
15
,
and
biophysical characterization of BioB and LipA continue, it will be fascinating to compare
- 36 -
and contrast how these two enzymes use different types of Fe-S clusters to catalyze
similar reactions.
1.6.5. Coproporphyrinogen III oxidase: a newfoldfor an Ado* enzyme
Coproporphyrinogen III oxidase (HemN) oxidatively decarboxylates propionate side
chains of coproporphyrinogen
III to vinyl groups153-156,an essential step in heme and
chlorophyll biosynthesis under certain growth conditions (Fig. 1.6f). The enzyme was
overexpressed and purified from E. coli1 56 and is known to contain a 4Fe-4S cluster,
ligated by the three Cys residues of the CxxxCxxC motif, and by AdoMet at a unique Fe
of the cluster 56 1' 57 . Surprisingly, the crystal structure of HemN 5 7 revealed not one, but
two AdoMet molecules bound to the enzyme (PDB code 1OLT; Fig. 1.17). Both of these
AdoMet sites were interpreted as physiologically relevant, but structural comparison with
BioB suggests that the second AdoMet molecule may be occupying the substrate-binding
site (discussed in chapter 5). Interestingly, the fold of HemN is not the same as that of
BioB, but the interactions of the 4Fe-4S cluster with AdoMet are almost identical
between the two enzymes (discussed in chapter 5).
-37-
1.7. Summary
AdoCbl- and AdoMet- dependent radical enzymes catalyze difficult reactions via radicalbased mechanisms.
Both families of enzymes must prevent aberrant free-radical
reactions while still maintaining proper reactivity for catalysis. Ado. is implicated in the
reactions of both enzyme families. The AdoMet radical enzymes require a catalytic [4Fe4S]l+ cluster in order to produce Ado., and reduction of the cluster to the 1+ oxidation
state requires a physiological reducing system, whereas AdoCbl-dependent enzymes
require homolytic cleave of the C-Co bond to produce Ado..
Several AdoCbl radical
enzymes have AdoMet radical enzyme analogs, such as the class II and class III RNRs
and 5,6-LAM/2,3-LAM.
This thesis addresses key questions in the field of Ado. enzymes, such as: What is the
structural basis for AdoMet binding by the AdoMet radical enzymes? What, if any,
structural relationship exists between the AdoMet radical enzymes and the AdoCbl
radical enzymes? How do Ado. enzymes prevent radical propagation in the absence of
substrate or effector?
Chapters 2 and 3 describe the crystallization, structure
determination, and structure analysis of the AdoMet radical enzyme BioB. Chapter 4
describes the crystallization, structure determination, and structure analysis of the
AdoCbl radical enzyme 5,6-LAM. Chapter 5 provides a few final thoughts on the
relationship between these two superamilies of Ado. radical-producing enzymes.
- 38 -
1.8. Figures
(b)
(a
NH2
I
(c)
NH
2
I~~~~~~~~~H
J~
04
OH
OH
5'-deoxyadenosyl
radical (Ado.)
NH2
S-adenosylmethionine
(AdoMet or SAM)
-O
Adenosylcobalamin (AdoCbl or vitamin B12)
Figure 1.1. Molecular structures of (a) AdoCbl, (b) Ado., and (c) AdoMet.
Ad
H
Na
NH2
tN
Ad
HOCJ
N
H
=
x
X
H/H
HO
AdH
.o~X H
Ad'H
·
-
.- .
'',
Figure 1.2. Simplified catalytic cycle of AdoCbl-dependent mutases. Carbon-cobalt
bond homolysis leads to the formation of Ado. and Cob(II). Ado. abstracts a hydrogen
atom from the substrate, initiating the 1,2 shift, which is simplified in this scheme. The
product-derived radical abstracts an H atom from AdoH, reforming the Ado./Cob(II)
couple, which recombine to regenerate the intact cofactor. The stereochemical course of
the reaction differs from enzyme to enzyme. Adapted from reference1 58
-39-
C
NH,+
-e+
5-, cO
k
o
H
H3NN -
(f)
H2N
N
HO
OH
(e)
AcyI-ACP/
oxO'
(d)
0N
OH
7
02C
F
OH
NH3
Figure 1.3. Selected AdoMet-dependent enzymatic reactions. The enzymes that carry
out these reactions are: (a) DAPA synthase26 (see also Fig. 1.4); (b) SAM-tRNA
ribosyltransferase-isomerase' 5 9 ; (c) 5'-deoxyfluoroadenosine
synthase 160; (d) 1aminocyclopropane-l-carboxylate
synthasel 61 ; (e) acylhomoserine lactone synthasel 6 2
(ACP = acyl carrier protein); (f) 3-(3-amino-3-carboxypropyl) uridine synthase 16 3 ; (g)
SAM decarboxylase 64; (h) Cyclopropane fatty acid synthase61 5.
reference' 6 6 .
- 40 -
Adapted from
+
H
0
,NH3
KAPA
synthase
PLP
o
Co 2 -
CoAS
L-Alanine
Pimeloyl-CoA
DAPA
synthase /
ielaroonic acid)
PLP
AdoMet
NH2
H2 N
C02
synthase
synthase
DAPA (7,8-diaminopelargonic acid)
Mg-A TP
C0 2
HN
NH
H
Biotin
synthase
DTB (Dethiobiotin)
AdoMet
o
NADPH
Flavodoxin
FNR
HN
HN
H
NH
NH
H
H
Biotin (Vitamin B8)
Figure 1.4. The biotin biosynthetic pathway. Enzymes in the pathway are in bold
typeface. Cofactors, co-substrates, and other enzymes involved in pathway reactions
(flavodoxin, flavodoxin-NADPH oxidoreductase [FNR]) are listed in italics.
-41 -
Class I: Carbon skeleton mutases
(C)
CoA-S C0 2
_
Co
-02
C
10
+NH3
-O'C
O
(b) CoA-
(a) CoA-&O
0
-0 2 C
O
H NH
OC"02
=
_-
Cy"i
Class II: Eliminases
(e)
H2,
9[)H+H
20
]
[>;OH
OH
A(
|OH
OH
HO
(f)
H
-
+ NH
OH
~
HC
,-,,,~,OH
O H
[
'
L'
AHO
+
%O0
OH
H
I
OH
J
+ H20
(h) PPPO
H2N
[HO, _
HO
NH3
,Base
OH
PPPO--
,Base
+
HO
H20
Class IIl:Aminomutases
NH
(i)
+
a~
+
NH
+3
(J)
NH+3N
+H3
NH3
C02
COz
H3N+
H3NV~~co2
NH3'
H3N
NH3
Reactions are sorted into
Figure 1.5. The AdoCbl-dependent enzymatic reactions.
2
2
general classes according to Buckel and Golding . The reactions are catalyzed by: (a)
MCM; (b) ICM; (c) GM; (d) MGM; (e) DDH; (f) GDH; (g) EAL; (h) class II RNR; (i)
5,6-LAM; (j) OAM. See text for references.
- 42 -
Sulfur insertase
(a)
J,
HN
(b)
HN
HN
NH
NH
H
0
0
H
H
4
S S
4
Mutase
NH3
(C)
H
N-
-C2
=
H3 N--C
oN
2
NH3
Activase
(d)
H
o
N
H
0
Lyase
(e)
0
O-NI
0O~
Oxidoreductase
(f)
Figure 1.6. Several moderately or well -characterized AdoMet radical enzyme reactions,
sorted into categories. The reactions are those of (a) BioB; (b) LipA; (c) 2,3-LAM; (d)
the activases of class III RNR and PFL; (e) SPL; (f) HemN. See text for references.
43 -
HC2
CO2
SAM
synthetase
ATP
AdoMet
or SAM
Nu:
(MethylNu-CH
,
trnnefrnoN
rr ur
rur
vr
3
uvv
S
NH3
C02
Met
J-
HS
SA H
Methionine
synthase
H0- NH3
CO2
HomocysteinE e
SAH hydrolase
Adenosine
Figure 1.7. The AdoMet cycle. Enzymes involved in recycling AdoMet are shown in
italics. The reaction most often associated with AdoMet, methylation of a nucleophile
with production of S-adenosylhomocysteine (SAH), is shown as part of the cycle.
Adapted from referencel 6 6 .
- 44 -
I1
i
,·
·
I Bl I
%
'U 1
'r ,,
i
i~~~~~~~
4)
Figure 1.8. Crystal structure of the MeCbl-binding domain of MS51 . (a) Ribbon
representation of the MeCbl-binding Rossmann domain. fl-strands are colored dark blue
and MeCbl is shown in a ball-and-sphere representation in pink. This coloring scheme is
used in all ribbon diagrams in this chapter. (b) Details of the base-off binding of MeCbl
to MS, showing the conserved residues of the DxHxxG...Sxl...GG sequence
characteristic of all base-off binding enzymes (see text). The coloring scheme for this
and all other Figures is: C, gray; 0, red; N, blue; P, green. The carbon atoms of MeCbl
67 .
are shown in black. All structural Figures were made with PyMo11
- 45 -
(a)
(b)
t_
k-,
I
ti
i"
I_
Ij I
"I
I
Figure 1.9. Crystal structure of MCM. (a) The MCM-Cbl-desulfo-CoA complex52 .
Desulfo-CoA is shown in a ball-and-stick representation in black. The substrate-binding
TIM barrel is shown in green. (b) The substrate-free MCM-AdoCbl complex 68 . The
adenosyl group is colored in red. In this structure, the TIM barrel is spliced open relative
to the MCM-Cbl-desulfo-CoA complex.
- 46 -
(a)
(I
tllt
:·r·
(C)
H143
[II>
+I
H-,
H3o
'
E171
H143
H143
l
H
E170
,0-
K'
', H '
H
,0
N
K
+'
/O
H
HN
E170
I N,>
K,+ '
K
H3C
H
H,, :' H
'04
+o o
H3C
H
Figure 1.10. Crystal structure of DDH76 . (a) Close-up view of APCbl, the substrate (1,2
propanediol), and a potassium ion (orange sphere) in the active site. (b) Ribbon diagram
of DDH showing the substrate-binding TIM barrel in green and the AdoCbl-binding
subunit in light blue, with the central fl-sheet in dark blue. The adeninylpentyl moiety of
APCbl is shown in red, Cbl in pink, the substrate in dark grey, and the potassium ion is
shown as an orange sphere. (c) A hypothetical mechanism for the 1,2 shift in DDH,
wherein the K+ ion is directly involved in catalysis (adapted from reference1 58).
-47 -
Figure 1.11. Crystal structure of GM in complex with substrate and AdoCbl80 . Only the
C2'-endo conformation of Ado (red) is shown. Gruber et al modeled a mixture of
substrate (L-glutamate, shown here in black) and product ([2S,3S]-3-methylaspartate, not
shown) into the active site. The TIM barrel is shown in green.
l
0e
Figure 1.12. Crystal structure of class II RNR6 0 . The AdoCbl-binding region is shown
in blue, with the central B-hairpin highlighted in dark blue. APCbl is bound in the base-
on configuration and shown in pink. The adeninylpentyl moiety is disordered and is not
shown in this Figure.
- 48 -
I
II
.IT,
Figure 1.13. Crystal structure of the substrate-free AdoCbl-PLP-5,6-LAM complex, a
major contribution to this thesis (discussed in chapter 4). The protein is an
a 2 32 tetramer;
only the o unit is shown. The AdoCbl binding Rossmann domain is colored in blue,
with the central $-sheet highlighted in dark blue, and the remainder of the protein is
colored green. Cbl is colored pink, AdoH is colored red, and PLP is colored black. 5,6LAM is unusual when compared to other AdoCbl-dependent mutases in that AdoCbl
does not bind directly atop the center of the TIM barrel.
- 49 -
2+
1+
2+
-
F
SFe
e
2+
K
0.-H 3 C-S
H
5'
1+
Figure. 1.14. Radical generation by the 4Fe-4S cluster and AdoMet in the AdoMet
radical enzymes. (a) A model for radical generation, based on Frey and coworkers'
spectroscopic studies of 2,3-LAM6° '0° 9. (b) A model of AdoMet interacting with the
reduced 4Fe-4S cluster of PFL activase, from Broderick and coworkers' pioneering
'
ENDOR experiments'2 5 26.
- 50-
*HzC--Ado
* H2C--Ado
Lys
C C
N
H
N
R
-L.ys
) -
V"CH
. OH
-203PO
,OH
-2'I",
..
~03PO'.
N
H'
CH
CH3
N.
H3 C -Ado
CH3
R
I
4
,coo'
C,
H"' C
N H
H3 C--Ado
H3C- Ado
*CH
H
C CCO
COO-
R
N
-20
OH
3 U
3
N
H.
~:OH
CH3
CH
OH
-203PO'H
.
.:
H+
ICOO'
"I
UPO
H
HC
2
-203 PO
R
HN+
CH 3
CH3
R = H3 N(CH 2 ) 3
Ado-CH3 = 5'-deoxyadenosine
Figure. 1.15. Mechanism proposed by Frey and coworkers for the 2,3-LAM reaction,
modified from reference 68 . The boxed intermediate has been observed by EPR and
ESEEM spectroscopy and is kinetically competent1 13'116
-51 -
1A'1
Figure 1.16. Crystal structure of BioB'3 7. Only one subunit of the BioB homodimer is
shown. The structure is a major contribution to this thesis. Fe-S clusters are shown as
spheres (Fe, brown; S, yellow), AdoMet (red) and substrate (DTB, black) are shown in
ball-and-stick representations. The central Bf-sheetof the TIM barrel is highlighted in
dark blue.
- 52 -
Figure 1.17. Crystal structure of HemN'57 . The 4Fe-4S cluster is shown as spheres (Fe,
brown; S, yellow). Two AdoMet molecules were observed in the structure; the one that
chelates a unique Fe of the cluster is shown in red and the second molecule is shown in
black. -strands are colored dark blue.
- 53 -
1.9. References
1.
2.
3.
4.
5.
Chanarin, I. Historical review: a history of pernicious anaemia. Br. J. Haematol.
111, 407-415 (2000).
Minot, G.R. & Murphy, W.P. Treatment of pernicious anemia by special diet. J.
Am. Med. Asoc. 14, 470-476 (1926).
Rickes, E.L., Brink, N.G., Koniuszky, F.R., Wood, T.R. & Folkers, K. Crystalline
Vitamin B1 2 . Science 107, 396-397 (1948).
Lester-Smith, E. Purification of the anti-pernicious anaemia factor from liver.
Nature 161, 638-639 (1948).
Ledley, F.D. & Rosenblatt, D.S. Mutations in mut methylmalonic acidemia:
clinical and enzymatic correlations. Hum. Mutat. 9, 1-6 (1997).
6.
Rosenblatt, D.S. & Fenton, W.A. in The metabolic and molecular bases of
inherited disease. (eds. Scriver, C.R., Beadet, A.L., Valle, D. & Sly, W.S.) 38973933 (McGraw-Hill, New York, 2001).
7.
Barker, H.A., Weissbach, H. & Smyth, R.D. A coenzyme containing
pseudovitamin B12 . Proc. Natl. Acad. Sci. 44, 1093-1097 (1958).
8.
Stadtman, E.R., Overath, P., Eggerer, H. & Lynen, F. The role of biotin and
vitamin B12 coenzyme in propionate metabolism. Biochem. Biophys. Res.
9.
Commun. 2, 1-7 (1960).
Eggerer, H., Overath, P., Lynen, F. & Stadtman, E.R. On the mechanism of the
cobamide coenzyme dependent isomerization of methylmalonyl CoA to succinyl
10.
CoA. J. Am. Chem. Soc. 82, 2643-2644 (1960).
Hodgkin, D.C. et al. Structure of Vitamin B 12 . Nature 178, 64-66 (1956).
11.
Lenhert, P.G. & Hodgkin, D.C. Structure of the 5,6-dimethyl-
12.
benzimidazolylcobamide coenzyme. Nature 192, 937-938 (1961).
Frey, P.A. & Abeles, R.H. The Role of the B 12 Coenzyme in the Conversion of
13.
1,2-Propanediol to Propionaldehyde. J. Biol. Chem. 241, 2732-2733 (1966).
Orme-Johnson, W.H., Beinert, H. & Blakley, R.L. Cobamides and ribonucleotide
reduction. XII. The electron paramagnetic resonance spectrum of "active
coenzyme B1 2 ." J. Biol. Chem. 249, 2338-2343 (1974).
14.
15.
16.
Babior, B.M., Moss, T.H., Orme-Johnson, W.H. & Beinert, H. The mechanism of
action of ethanolamine ammonia-lyase, a B-12-dependent enzyme. The
participation of paramagnetic species in the catalytic deamination of 2aminopropanol. J. Biol. Chem. 249, 4537-4544 (1974).
Babior, B.M., Moss, T.H. & Gould, D.C. The mechanism of action of
ethanolamine ammonia lyase, a B12-dependent enzyme. A study of the reaction
by electron spin resonance spectrometry. J. Biol. Chem. 247, 4389-4392 (1972).
Banerjee, R. Radical peregrinations catalyzed by coenzyme B12-dependent
enzymes. Biochemistry 40, 6191-6198 (2001).
17.
Banerjee, R. Radical carbon skeleton rearrangements: catalysis by coenzyme
B12-dependent mutases. Chem. Rev. 103, 2083-2094 (2003).
18.
19.
Banerjee, R. & Ragsdale, S.W. The many faces of vitamin B12: catalysis by
cobalamin-dependent enzymes. Annu. Rev. Biochem. 72, 209-247 (2003).
Frey, P.A. Radical Mechanisms of Enzymatic Catalysis. Annu. Rev. Biochem. 70,
121-148 (2001).
- 54 -
20.
Frey, P.A. & Reed, G.H. Radical Mechanisms in Adenosylmethionine- and
Adenosylcobalamin-Dependent Enzymatic Reactions. Arch. Biochem. Biophys.
382, 6-14 (2000).
21.
22.
23.
Toraya, T. Enzymatic radical catalysis: coenzyme B 12 -dependent diol
dehydratase. Chem. Rec. 2, 352-366 (2002).
Buckel, W. & Golding, B.T. Glutamate and 2-methyleneglutarate mutase: from
microbial curiosities to paradigms for coenzyme B 12 -dependent enzymes. Chem.
Soc. Rev. 25, 329-337 (1996).
Stubbe, J., Licht, S., Gerfen, G.J., Silva, D. & Booker, S.J. Adenosylcobalamin-
dependent ribonucleotide reductases: still amazing but no longer confusing. in
Vitamin B12 and B12-proteins(eds. Krautler, B., Arigoni, D. & Golding, B.T.)
321-331 (Wiley-VCH, Weinheim, 1998).
24.
25.
26.
27.
Moss, M. & Frey, P.A. The role of S-adenosylmethionine in the lysine 2,3aminomutase reaction. J. Biol. Chem. 262, 14859-14862 (1987).
Chirpich, T.P., Zappia, V., Costilow, R.N. & Barker, H.A. Lysine 2,3aminomutase. Purification and properties of a pyridoxal phosphate and Sadenosylmethionine-activated enzyme. J. Biol. Chem. 245, 1778-1789 (1970).
Stoner, G.L. & Eisenberg, M.A. Biosynthesis of 7, 8-diaminopelargonic acid from
7-keto-8-aminopelargonic acid and S-adenosyl-L-methionine. The kinetics of the
reaction. J. Biol. Chem. 250, 4037-4043 (1975).
Zhang, S., Sanyal, I., Bulboaca, G.H. & Rich, A. The gene for biotin synthase
from Saccharomyces cerevisiae: cloning, sequencing, and complementation of
Escherichia coli strains lacking biotin synthase. Arch. Biochem. Biophys. 309, 2935 (1994).
28.
29.
Baldet, P., Gerbling, H., Axiotis, S. & Douce, R. Biotin biosynthesis in higher
plant cells. Identification of intermediates. Eur. J. Biochem. 217, 479-485 (1993).
Sanyal, I., Cohen, G. & Flint, D.H. Biotin synthase: purification, characterization
as a [2Fe-2S]cluster protein, and in vitro activity of the Escherichia coli bioB
gene product. Biochemistry 33(1994).
30.
Escalletes, F., Florentin, D., Tse Sum Bui, B., Lesage, D. & Marquet, A. Biotin
Synthase Mechanism: Evidence for Hydrogen Transfer from the Substrate into
Deoxyadenosine. J. Am. Chem. Soc. 121, 3571-3578 (1999).
31.
Sofia, H.J., Chen, G., Hetzler, B.G., Reyes-Spindola, J.F. & Miller, N.E. Radical
SAM, a novel protein superfamily linking unresolved steps in familiar
biosynthetic pathways with radical mechanisms: functional characterization using
new analysis and information visualization methods. Nucleic Acids Res. 29, 10971106 (2001).
32.
33.
34.
35.
Beinert, H., Holm, R.H. & Munck, E. Iron-sulfur clusters: nature's modular,
multipurpose structures. Science 277, 653-659 (1997).
Huber, C. & Wachtershrauwer, G. Activated acetic acid by carbon fixation on
(Fe,Ni)S under primordial conditions. Science 276, 245-247 (1997).
Huber, C. & Wachtershrauwer, G. Peptides by activation of amino acids with CO
on (Ni,Fe)S surfaces: implications for the origin of life. Science 281, 670-672
(1998).
Scott, A.I. & Roessner, C.A. Biosynthesis of cobalamin (vitamin B 12 ). Biochem.
Soc. Trans. 4, 613-620 (2002).
- 55 -
36.
Torrents, E., Aloy, P., Gibert, I. & Rodriquez-Trelles, F. Ribonucleotide
Reductases: Divergent Evolution of an Ancient Enzyme. J. Mol. Evol. 55, 138152 (2002).
37.
Reichard, P. The evolution of ribonucleotide reduction. Trends Biochem. Sci. 22,
38.
Eschenmoser, A. Vitamin B12: Experiments concerning the origin of its molecular
39.
Hesse, H. & Hoefgen, R. Molecular aspects of methionine biosynthesis. Trends
81-85 (1997).
structure. Angew. Chem. Int. Ed. 27, 5-39 (1988).
40.
41.
Plant. Sci. 6, 259-262 (2003).
Licht, S., Gerfen, G.J. & Stubbe, J. Thiyl Radicals in Ribonucleotide Reductases.
Science 271, 477-481 (1996).
Gerfen, G.J., Licht, S., Willems, J.-P., Hoffinan, B.M. & Stubbe, J. electron
paramagnetic resonance investigations of a kinetically competent intermediate
formed in ribonucleotide reduction: evidence for a thiyl radical-Cob(II)alamin
interaction. J. Am. Chem. Soc. 118, 8192-8197 (1996).
42.
43.
Licht, S.S., Lawrence, C.C. & Stubbe, J. Thermodynamic and kinetic studies on
carbon-cobalt bond homolysis by ribonucleoside triphosphate reductase: the
importance of entropy in catalysis. Biochemistry 38, 1234 (1999).
Brown, K.L. & Li, J. Activation parameters for the carbon-cobalt bond homolysis
of coenzyme B 12 induced by the B12-dependent ribonucleotide reductase from
Lactobacillus leichmannii. J. Am. Chem. Soc. 120, 9466-9474 (1998).
44.
Cantoni, G.L. Methylation of nicotinamide with soluble enzyme system from rat
liver. J. Biol. Chem. 189, 203-216 (1951).
45.
Chowdhury, S. & Banerjee, R. Thermodynamic and kinetic characterization of
Co-C bond homolysis catalyzed by coenzyme B(12)-dependent methylmalonylCoA mutase. Biochemistry 39, 7998-8006 (2000).
46.
47.
Marsh, E.N. & P., B.D. Coupling of cobalt-carbon bond homolysis and hydrogen
atom abstraction in adenosylcobalamin-dependent glutamate mutase.
Biochemistry 37, 11864-11872 (1998).
Murzin, A.G., Brenner, S.E., Hubbard, T. & Chothia, C. SCOP: a structural
50.
classification of proteins database for the investigation of sequences and
structures. J. Mol. Biol. 247, 536-540 (1995).
Nagano, N., Orengo, C.A. & Thornton, J.M. One fold with many functions: The
evolutionary relationships between TIM barrel families based on their sequences,
structures and functions. J. Mol. Biol. 321, 741-765 (2002).
Copley, R.R. & Bork, P. Homology among (a)8 barrels: Implications for the
evolution of metabolic pathways. J. Mol. Biol. 303, 627-641 (2000).
Hennig, M. et al. A TIM barrel protein without enzymatic activity? Crystal-
51.
structure of narbonin at 1.8 A resolution. FEBS Lett. 306, 80-84 (1992).
Drennan, C.L., Huang, S., Drummond, J.T., Matthews, R.G. & Ludwig, M.L.
48.
49.
How a protein binds B12: A 3.0 A X-ray structure of B12 -binding domains of
52.
methionine synthase. Science 266, 1669-1674 (1994).
Mancia, F. et al. How coenzyme B 12 radicals are generated: the crystal structure
of methylmalonyl-coenzyme A mutase at 2 A resolution. Structure 4, 339-350
(1996).
- 56 -
53.
54.
55.
Reitzer, R. et al. Glutamate mutase from Clostridium cochlearium: the structure
of a coenzyme B12-dependent enzyme provides new mechanistic insights.
Structure Fold. Des. 7, 891-902 (1999).
Chang, C.H. & Frey, P.A. Cloning, sequencing, heterologous expression,
purification, and characterization of adenosylcobalamin-dependent D-lysine 5, 6aminomutase from Clostridium sticklandii. J. Biol. Chem. 275, 106-114 (2000).
Ratnatilleke, A., Vrijbloed, J.W. & Robinson, J.A. Cloning and sequencing of the
coenzyme B(12)-binding domain of isobutyryl-CoA mutase from Streptomyces
cinnamonensis, reconstitution of mutase activity, and characterization of the
recombinant enzyme produced in Escherichia coli. J. Biol. Chem. 274, 3167931685 (1999).
56.
57.
Beatrix, B., Zelder, O., Liner, D. & Buckel, W. Cloning, sequencing and
expression of the gene encoding the coenzyme B 12-dependent 2methyleneglutarate mutase from Clostridium barkeri in Escherichia coli. Eur. J.
Biochem. 221, 101-109 (1994).
Chen, H.P., Wu, S.H., Lin, Y.L., Chen, C.M. & Tsay, S.S. Cloning, sequencing,
heterologous expression, purification, and characterization of adenosylcobalamindependent D-ornithine aminomutase from Clostridium sticklandii. J. Biol. Chem.
58.
276, 44744-44750 (2001).
Masuda, J., Shibata, N., Morimoto, Y., Toraya, T. & Yasuoka, N. How a protein
generates a catalytic radical from coenzyme B12 : X-ray structure of a dioldehydratase-adeninylpentylcobalamin complex. Structure Fold. Des. 8, 775
(2000).
59.
60.
61.
62.
Abend, A. et al. Ethanolamine ammonia-lyase has a "base-on" binding mode for
coenzyme B12 . Arch. Biochem. Biophys. 370, 138-141 (1999).
Sintchak, M.D., Arjara, G., Kellog, B.A., Stubbe, J. & Drennan, C.L. The crystal
structure of class II ribonucleotide reductase reveals how an allosterically
regulated monomer mimics a dimer. Nat. Struct. Biol., 293-300 (2002).
Borths, E.L., Locher, K.P., Lee, A.T. & Rees, D.C. The structure of Escherichia
coli BtuF and binding to its cognate ATP binding cassette transporter. Proc. Natl.
Acad. Sci. 99, 16642-16647 (2002).
Bauer, C.B. et al. Three-dimensional structure of ATP:corrinoid
adenosyltransferase from Salmonella typhimurium in its free state, complexed
with MgATP, or complexed with hydroxycobalamin and MgATP. Biochemistry
63.
40, 361-374 (2001).
Ouyang, L., Rulis, P., Ching, W.Y., Nardin, G. & Randaccio, L. Accurate
redetermination of the X-ray structure and electronic bonding in
adenosylcobalamin. Inorg. Chem. 43, 1235-1241 (2004).
64.
Ng, F..T.T. & Rempel, G.L. Ligand effects on transition metal-alkyl bond
dissociation energies. J. Am. Chem. Soc. 104, 621-623 (1982).
65.
Geno, M.K. & Halpern, J. Why does nature not use the porphyrin ligand in
vitamin B1 2? J. Am. Chem. Soc. 109, 1238-1240 (1987).
66.
Champloy, F. et al. EXAFS data indicate a "normal" axial cobalt-nitrogen bond of
the organo-B 12 cofactor in the two coenzyme B 12-dependent enzymes glutamate
mutase and 2-methyleneglutarate mutase. J. Am. Chem. Soc. 121, 11780-11789
(1999).
- 57 -
67.
Scheuring, E., Padmakumar, R., Banerjee, R. & Chance, M.R. Extended X-ray
absorption fine structure analysis of coenzyme B 12 bound to methylmalonylcoenzyme A mutase using global mapping techniques. J. Am. Chem. Soc. 119,
12192-12200 (1997).
68.
69.
Mancia, F. & Evans, P.R. Conformational changes on substrate binding to
methylmalonyl CoA mutase and new insights into the free radical mechanism.
Structure 6, 711-720 (1998).
Smith, D.M., Golding, B.T. & Random, L. Understanding the mechanism of B12 dependent methylmalonyl-CoA mutase: partial proton transfer in action. J. Am.
Chem. Soc. 121, 9388-9399 (1999).
70.
Maiti, N., Widjaja, L. & Banerjee, R. Proton transfer from histidine 244 may
facilitate the 1,2 rearrangement reaction in coenzyme B(12)-dependent
methylmalonyl-CoA mutase. J. Biol. Chem. 274, 32733-32737 (1999).
71.
Thoma, N.H., Evans, P.R. & Leadlay, P.F. Protection of radical intermediates at
the active site of adenosylcobalamin-dependent methylmalonyl-CoA mutase.
Biochemistry 39, 9213-9221 (2000).
72.
73.
Mancia, F., Smith, G.A. & Evans, P.R. Crystal structure of substrate complexes of
methylmalonyl-CoA mutase. Biochemistry 25, 7999-8005 (1999).
Vlasie, M.D. & Banerjee, R. Tyrosine 89 accelerates Co-carbon bond homolysis
in methylmalonyl-CoA mutase. J. Am. Chem. Soc. 125, 5431-5435 (2003).
74.
Lee, H.A.J. & Abeles, R.H. Purification and properties of dioldehydrase, and
enzyme requiring a cobamide coenzyme. J. Biol. Chem. 238, 2367-2373 (1963).
75.
Toraya, T. Radical catalysis in coenzyme B 12 -dependent isomerization
(eliminating) reactions. Chem. Rev. 103, 2095-2127 (2003).
76.
Shibata, N. et al. A new mode of B 12 binding and the direct participation of a
potassium ion in enzyme catalysis: X-ray structure of diol dehydratase. Structure
7, 997-1008 (1999).
77.
78.
Liao, D.-I., Dotson, G., Turner, I.J., Reiss, L. & Emptage, M. Crystal structure of
substrate free form of glycerol dehydratase. J. Inorg. Biochem. 93, 84-91 (2003).
Shibata, N., Masuda, J., Morimoto, Y., Yasuoka, N. & Toraya, T. Substrateinduced conformational change of a coenzyme Bl2-dependent enzyme: crystal
structure of the substrate-free form of diol dehydratase. Biochemistry 41, 1260712617 (2002).
79.
80.
Barker, H.A., Smyth, R.D., Wilson, R.M. & Weissbach, H. The purification and
properties of P-methylaspartase. J. Biol. Chem. 234, 320-328 (1959).
Gruber, K., Reitzer, R. & Kratky, G. Radical Shuttling in a Protein: Ribose
Pseudorotation Controls Alkyl-Radical Transfer in the Coenzyme B(12)
Dependent Enzyme Glutamate Mutase. Angew. Chem. Int. Ed. 40, 3377-3380
(2001).
81.
82.
Bothe, H. et al. Identification of the 4-glutamyl radical as an intermediate in the
carbon skeleton rearrangement catalyzed by coenzyme Bl2-dependent glutamate
mutase from Clostridium cochlearium. Biochemistry 37, 4105-4113 (1998).
Hoffman, B., Konrat, R., Bothe, H., Buckel, W. & Krautler, B. Structure and
dynamics of the Bi2 -binding subunit of glutamate mutase from Clostridium
cochlearium. Eur. J. Biochem. 263, 178-188 (1999).
- 58 -
83.
Tollinger, M., Konrat, R., Hilbert, B.H., Marsh, E.N. & Krautler, B. How a
protein prepares for B12 binding: structure and dynamics of the Bl2-binding
subunit of glutamate mutase from Clostridium tetanomorphum. Structure 6, 10211033 (1998).
84.
Hoffman, B. et al. A protein pre-organized to trap the nucleotide moiety of
coenzyme Bl2: refined solution structure of the B12-binding subunit of glutamate
mutase from Clostridium tetanomorphum. Chembiochem 2, 643-655 (2001).
85.
Tollinger, M. et al. The B 12 -binding subunit of glutamate mutase from
Clostridium tetanomorphum traps the nucleotide moiety of coenzyme B1 2 . J. Mol.
Biol. 309(2001).
86.
87.
88.
89.
Booker, S.J., Licht, S., Broderick, J.B. & Stubbe, J. Coenzyme B12-dependent
ribonucleotide reductase: evidence for the participation of five cysteine residues
in ribonucleotide reduction. Biochemistry 33(1994).
Holmgreen. Thioredoxin. Annu. Rev. Biochem. 54, 205-235 (1985).
Abeles, R.H. & Beck, W.S. The mechanism of action of cobamide coenzyme in
the ribonucleotide reductase reaction. J. Biol. Chem. 242, 3589-3593 (1967).
Hogenkamp, H.P., Ghambeer, R.K., Brownson, C., Blakley, R.L. & Vitols, E.
Cobamides and ribonucleotide reduction. VI. Enzyme-catalyzed hydrogen
exchange between water and deoxyadenosylcobalamin. J. Biol. Chem. 243, 799808 (1968).
90.
Licht, S.S., Booker, S.J. & Stubbe, J. Studies on the catalysis of carbon-cobalt
bond homolysis by ribonucleoside triphosphate reductase: evidence for concerted
carbon-cobalt bond homolysis and thiyl radical formation. Biochemistry 38, 1221-
91.
1233 (1999).
Lawrence, C.C., Gerfen, G.J., Samano, V., Nitsche, R. & Robins, M.J. Binding of
92.
Cob(II)alamin to the adenosylcobalamin-dependent ribonucleotide reductase from
Lactobacillus leichmannii. Identification of dimethylbenzimidazole as the axial
ligand. J. Biol. Chem. 274(1999).
Jordan, A. & Reichard, P. Ribonucleotide reductases. Annu. Rev. Biochem. 67,
71-98 (1998).
93.
Tamao, Y. & Blakley, R.L. Direct spectrophotometric observation of an
intermediate formed from deoxyadenosylcobalamin in ribonucleotide reduction.
Biochemistry 12, 24-34 (1973).
94.
Uhlin, U. & Eklund, H. Structure of ribonucleotide reductase protein R1. Nature
370(1994).
95.
Logan, D.T., Andersson, J., Sjoberg, B.M. & Nordlund, P. A glycyl radical site in
the crystal structure of a class III ribonucleotide reductase. Science 283, 149996.
1504 (199).
Baker, J.J., van der Drift, C. & Stadtman, T.C. Purification and properties of P-
lysine mutase, a pyridoxal phosphate and B 12 coenzyme dependent enzyme.
Biochemistry 12, 1054-1063 (1973).
97.
Stadtman, T.C. Anaerobic degradation of lysine. II. Cofactor requirements and
properties of the soluble enzyme system. J. Biol. Chem. 238, 2766-2773 (1963).
98.
Stadtman, T.C. & Renz, P. The B 12 -coenzyme dependent step in lysine
fermentation: the migration of an amino group from carbon 6 to carbon 5 forming
3,5-diaminohexanoate.
Fed. Proc. 26, 343 (1967).
- 59 -
99.
100.
Dekker, E.E. & Barker, H.A. Identification and Cobamide Coenzyme-dependent
Formation of 3,5-Diaminohexanoic Acid, and Intermediate in Lysine
Fermentation. J. Biol. Chem. 243, 3232-3237 (1968).
Morley, C.G.D. & Stadtman, T.C. Studies on the fermentation of D-a-Lysine.
Purification and properties of an adenosine triphosphate regulated B12-coenzymedependent D-a-Lysine mutase complex from Clostridium sticklandii.
Biochemistry 9, 1970 (1970).
101.
Morley, C.G.D. & Stadtman, T.C. The role of pyridoxal phosphate in the B12
coenzyme-dependent D-a-Lysine mutase reaction. Biochemistry 11, 600-605
102.
Barker, H.A., Kahn, J.M. & Hedrick, L. Pathway of lysine degradation in
Fusobacterium nucleatum. J. Bact. 152, 201-207 (1982).
103.
Tang, K.H., Chang, C.H. & Frey, P.A. Electron transfer in the substrate-
(1972).
dependent suicide inactivation of Lysine 5,6-Aminomutase. Biochemistry 40,
5190-5199 (2001).
104.
Danen, W.C. & West, C.T. An electron spin resonance investigation of the 1Aziridylcarbinyl and related free radicals. J. Am. Chem. Soc. 96, 2447-2453
(1973).
105.
Song, K.B. & Frey, P.A. Molecular properties of lysine-2,3-aminomutase. J. Biol.
106.
Chem. 266, 7651-7655 (1991).
Cosper, N.J., Booker, S.J., Ruzicka, F., Frey, P.A. & Scott, R.A. Direct FeS
cluster involvement in generation of a radical in lysine 2,3-aminomutase.
Biochemistry 39, 15668-15673 (2000).
107.
Baraniak, J., Moss, M.L. & Frey, P.A. Lysine 2,3-aminomutase. Support for a
mechanism of hydrogen transfer involving S-adenosylmethionine. J. Biol. Chem.
108.
Moss, M.L. & Frey, P.A. Activation of lysine 2,3-aminomutase by S-
109.
adenosylmethionine. J. Biol. Chem. 265, 18112-18115 (1990).
Chen, D., Walsby, C.J., Hoffman, B.M. & Frey, P.A. Coordination and
264, 1357-1360 (1989).
mechanism of reversible cleavage of S-adenosylmethionine by the [4Fe-4S]
110.
center in lysine 2,3-aminomutase. J. Am. Chem. Soc. 125, 11788-11789 (2003).
Petrovich, R.M., Ruzicka, F.J., Reed, G.H. & Frey, P.A. Characterization of iron-
sulfur clusters in lysine 2,3-aminomutase by electron paramagnetic resonance
111.
spectroscopy. Biochemistry 31, 10774-10781 (1992).
Lieder, K.W., Booker, S.J., Ruzicka, F.J., Beinert, H. & Reed, G.H. S-
112.
Adenosylmethionine-dependent reduction of lysine 2,3-aminomutase and
observation of the catalytically functional iron-sulfur centers by electron
paramagnetic resonance. Biochemistry 37, 2578-2585 (1998).
Han, O. & Frey, P.A. Chemical model for the pyridoxal 5'-phosphate dependent
lysine aminomutases. J. Am. Chem. Soc. 112, 8982-8983 (1990).
113.
Ballinger, M.D., Frey, P.A. & Reed, G.H. Structure of a substrate radical
114.
intermediate in the reaction of lysine 2,3-aminomutase. Biochemistry 31, 1078210789 (1992).
Ballinger, M.D., Frey, P.A., Reed, G.H. & LoBrutto, R. Pulsed electron
paramagnetic resonance studies of the lysine 2,3-aminomutase substrate radical:
- 60 -
evidence for participation of pyridoxal 5'-phosphate in a radical rearrangement.
Biochemistry 34, 10086-10093 (1995).
115.
116.
117.
Ballinger, M.D., Reed, G.H. & Frey, P.A. An organic radical in the lysine 2,3aminomutase reaction. Biochemistry 31, 949-953 (1992).
Chang, C.H., Ballinger, M.D., Reed, G.H. & Frey, P.A. Lysine 2,3-Aminomutase:
Rapid Mix-Freeze-Quench Electron Paramagnetic Resonance Studies
Establishing the Kinetic Competence of a Substrate-Based Radical Intermediate.
Biochemistry 35, 11081-11084 (1996).
Knappe, J. & Schmitt, T. A novel reaction of S-adenosyl-L-methionine correlated
with the activation of pyruvate formate-lyase. Biochem. Biophys. Res. Commun.
23, 1110-1117 (1976).
118.
119.
Knappe, J., Blaschkowski, H.P., Grobner, P. & Schmitt, T. Pyruvate formatelyase of Escherichia coli: the acetyl-enzyme intermediate. Eur. J. Biochem. 50,
253-263 (1974).
Knappe, J. et al. Pyruvate formate-lyase reaction in Escherichia coli. The
enzymatic system converting an inactive form of the lyase into the catalytically
active enzyme. Eur. J. Biochem. 11, 316-327 (1969).
120.
Broderick, J.B. et al. Pyruvate formate-lyase activating enzyme is an iron-sulfur
protein. J. Am. Chem. Soc. 119, 7396-7397 (1997).
121.
Kulzer, R., Pils, T., Kappl, R., Huttermann, J. & Knappe, J. Reconstitution and
characterization of the polynuclear iron-sulfur cluster in pyruvate formate-lyaseactivating enzyme. Molecular properties of the holoenzyme form. J. Biol. Chem.
273, 4897-4903 (1998).
122.
123.
Henshaw, T.F., Cheek, J. & Broderick, J.B. The [4Fe-4S]1+ cluster of pyruvate
formate-lyase activating enzyme generates the glycyl radical on pyruvate formatelyase: EPR-detected single turnover. J. Am. Chem. Soc. 122, 8331-8332 (2000).
Conradt, H., Hohmann-Berger, M., Hohmann, H.P., Blaschkowski, H.P. &
Knappe, J. Pyruvate formate-lyase (inactive form) and pyruvate formate-lyase
activating enzyme of Escherichia coli: isolation and structural properties. Arch.
Biochem. Biophys. 228, 133-142 (1984).
124.
Frey, M., Rothe, M., Wagner, A.F.V. & Knappe, J. Adenosylmethioninedependent synthesis of the glycyl radical in pyruvate formate-lyase by abstraction
of the glycine C-2 pro-S hydrogen atom. Studies of [2 H]glycine-substituted
enzyme and peptides homologous to the glycine 734 site. J. Biol. Chem. 269,
12432-12437 (1994).
125.
Krebs, C., Broderick, W.E., Henshaw, T.F., Broderick, J.B. & Huynh, B.H.
Coordination of adenosylmethionine to a unique iron site of the [4Fe-4S] of
pyruvate formate-lyase activating enzyme: A mossbauer spectroscopic study. J.
Am. Chem. Soc. 124, 912-913 (2001).
126.
127.
Walsby, C.J., Ortillo, D., Broderick, W.E., Broderick, J.B. & Hoffman, B.M. An
anchoring role for FeS clusters: chelation of the amino acid moiety of Sadenosylmethionine to the unique iron site of the [4Fe-4S] cluster of pyruvate
formate-lyase activating enzyme. J. Am. Chem. Soc. 124, 11270-11271 (2002).
Fontecave, M., Eliasson, R. & Reichard, P. Oxygen-sensitive ribonucleoside
triphosphate reductase is present in anaerobic Escherichia coli. Proc. Natl. Acad.
Sci. 86, 2147-2151 (1989).
-61 -
128.
129.
130.
131.
Harder, J., Eliason, R., Pontis, E., Ballinger, M.D. & Reichard, P. Activation of
the anaerobic ribonucleotide reductase from Escherichia coli by Sadenosylmethionine. J. Biol. Chem. 267, 25548-25552 (1992).
Ollagnier, S. et al. Activation of the anaerobic ribonucleotide reductase from
Escherichia coli. The essential role of the iron-sulfur center for Sadenosylmethionine reduction. J. Biol. Chem. 272, 24216-24223 (1997).
Padovani, D., Thomas, F., Trautwein, A.X., Mulliez, E. & Fontecave, M.
Activation of class III ribonucleotide reductase from E. coli. The electron transfer
from the iron-sulfur center to S-adenosylmethionine. Biochemistry 40, 6713-6719
(2001).
Tamarit, J., Mulliez, E., Meier, C., Trautwein, A.X. & Fontecave, M. The
anaerobic ribonucleotide reductase from Escherichia coli. The small protein is an
activating enzyme containing a [4fe-4s](2+) center. J. Biol. Chem. 274, 31291-
132.
133.
31296 (1999).
Tamarit, J. et al. The activating component of the anaerobic ribonucleotide
reductase from Escherichia coli. An iron-sulfur center with only three cysteines.
J. Biol. Chem. 275, 15669-15675 (2000).
Mulliez, E., Padovani, D., Atta, M., Alcouffe, C. & Fontecave, M. Activation of
class III ribonucleotide reductase by flavodoxin: A protein radical-driven electron
transfer to the iron-sulfur center. Biochemistry 40, 3730-3736 (2001).
134.
135.
136.
137.
Ohsawa, I. et al. Cloning of the biotin synthetase gene from Bacillus sphaericus
and expression in Escherichia coli and Bacilli. Gene 80, 39-48 (1989).
Mejean, A. et al. Highly purified biotin synthase can transform dethiobiotin into
biotin in the absence of any other protein, in the presence of photoreduced
deazaflavin. Biochem. Biophys. Res. Commun. 217, 1231-1237 (1995).
Ugulava, N.B., Gibney, B.R. & Jarrett, J.T. Biotin synthase contains two distinct
iron-sulfur cluster binding sites: chemical and spectroelectrochemical analysis of
iron-sulfur interconversions. Biochemistry 40, 8343-8351 (2001).
Berkovtich, F., Nicolet, Y., Wan, J.T., Jarrett, J.T. & Drennan, C.L. Crystal
structure of biotin synthase, an S-adenosylmethionine-dependent radical enzyme.
Science 303, 76-79 (2004).
138.
Cosper, M.M. et al. The [4Fe-4S]2 + cluster in reconstituted biotin synthase binds
S-adenosylmethionine. J. Am. Chem. Soc. 124, 14006-14007 (2002).
139.
Cosper, M.M. et al. Structural studies of the interaction of S-adenosylmethionine
with the [4Fe-4S] clusters in biotin synthase and pyruvate formate-lyase
activating enzyme. Protein Sci. 12, 1573-1577 (2003).
140.
Tse Sum Bui, B. et al. Fate of the (2Fe-2S)2+ cluster of Escherichia coli biotin
synthase during reaction: a Mossbauer characterization. Biochemistry 42, 87918798 (2003).
141.
142.
Ugulava, N.B., Sacanell, C.J. & Jarrett, J.T. Spectroscopic Changes during a
Single Turnover of Biotin Synthase: Destruction of a [2Fe-2S] Sluster
Accompanies Sulfur Insertion. Biochemistry 40, 8352-8358 (2001).
Tse Sum Bui, B. et al. Biotin synthase mechanism: on the origin of sulphur. FEBS
Lett. 440, 226-230 (1998).
- 62 -
143.
144.
Jameson, G.N., Cosper, M.M., Hernandez, H.L., Johnson, M.K. & Huynh, B.H.
Role of the [2Fe-2S] cluster in recombinant Escherichia coli biotin synthase.
Biochemistry 43, 2022-2031 (2004).
Ollagnier-de Choudens, S., Mulliez, E. & Fontecave, M. The PLP-dependent
biotin synthase from Escherichia coli: mechanistic studies. FEBS Lett. 532, 465468 (2002).
145.
Ollagnier-de Choudens, S., Mulliez, E., Hewiston, K.S. & Fontecave, M. Biotin
synthase is a pyridoxal phosphate-dependent cysteine desulfurase. Biochemistry
41, 9145-9152 (2002).
146.
147.
148.
Picciocchi, A., Douce, R. & Alban, C. Biochemical characterization of the
Arabidopsis biotin synthase reaction. The importance of mitochondria in biotin
synthesis. Plant Physiol. 127, 1224-1233 (2001).
Reed, L.J. A trail of research from lipoic acid to alpha-keto acid dehydrogenase
complexes. J. Biol. Chem. 276, 38329-38336 (2001).
Zhao, X., Miller, J.R., Jiang, Y., Marletta, M.A. & Cronan, J.E. Assembly of the
covalent linkage between lipoic acid and its cognate enzymes. Chem. Biol. 10,
1293-1302 (2003).
149.
150.
Cicchillo, R.M. et al. Lipoyl synthase requires two equivalents of S-adenosyl-Lmethionine to synthesize one equivalent of lipoic acid. Biochemistry 43, 63786386 (2004).
Miller, J.R. et al. Escherichia coli LipA is a lipoyl synthase: in vitro biosynthesis
of lipoylated pyruvate dehydrogenase complex from octanoyl-acyl carrier protein.
Biochemistry 39, 15166-15178 (2000).
151.
152.
Cicchillo, R.M. et al. Escherichia coli lipoyl synthase binds two distinct [4Fe-4S]
clusters per polypeptide. Biochemistry ASAP(2004).
Trainor, D.A., Parry, R.J. & Gitterman, A. Biotin Biosynthesis. 2.
Stereochemistry of Sulfur Introduction at C-4 of Dethiobiotin. J. Am. Chem. Soc.
102, 1467-1468 (1980).
153.
Keithly, J.H. & Nadler, K.D. Protoporphyrin formation in Rhizobiumjaponicum.
J. Bact. 154, 838-845 (1983).
154.
Poulson, R. & Polglase, W.J. Aerobic and anaerobic coproporphyrinogenase
activities in extracts from Saccharomyces cerevisiae. Purification and
characterization. J. Biol. Chem. 249, 6367-6371 (1974).
155.
Tait, G.H. Coproporphyrinogenase activity in extracts from Rhodopseudomonas
spheroides. Biochem. Biophys. Res. Commun. 37, 116-122 (1969).
156.
Layer, G., Verfurth, K., Mahlitz, E. & Jahn, D. Oxygen-independent
coproporphyrinogen-III oxidase HemN from Escherichia coli. J. Biol. Chem. 277,
34136-34142 (2002).
157.
158.
159.
Layer, G., Moser, J., Heinz, D.W., Jahn, D. & Schubert, W.D. Crystal structure of
coproporphyrinogen III oxidase reveals cofactor geometry of Radical SAM
enzymes. EMBO J. 22, 6214-6224 (2003).
Marsh, E.N. & Drennan, C.L. Adenosylcobalamin-dependent isomerases: new
insights into structure and mechanism. Curr. Opin. Chem. Biol. 5, 499-505
(2001).
Iwata-Reuyl, D. Biosynthesis of the 7-deazaguanosine hypermodified nucleosides
of transfer RNA. Bioorg. Chem. 31, 24-43 (2003).
- 63 -
160.
O'Hagan, D., Schaffrath, C., Cobb, S.L., Hamilton, J.T. & Murphy, C.D.
Biosynthesis of an organofluorine molecule. Nature 416, 279 (2002).
161.
White, M.F., Vasquez, J., Yang, S.F. & Kirsch, J.F. Expression of apple 1-
aminocyclopropane-1-carboxylate synthase in Escherichia coli: kinetic
characterization of wild-type and active-site mutant forms. Proc. Natl. Acad. Sci.
91, 12428-12432 (1994).
162.
163.
Val, D.L. & Cronan, J.E.J. In vivo evidence that S-adenosylmethionine and fatty
acid synthesis intermediates are the substrates for the LuxI family of autoinducer
synthases. J. Bact. 10, 2644-2651 (1998).
Nishimura, S., Taya, Y., Kuchino, Y. & Oashi, Z. Enzymatic synthesis of 3-(3-
amino-3-carboxypropyl)uridine in Escherichia coli phenylalanine transfer RNA:
transfer of the 3-amino-acid-3-carboxypropyl group from S-adenosylmethionine.
164.
165.
Biochem. Biophys. Res. Commun. 57, 702-708 (1974).
Tabor, C.W. & Tabor, H. Polyamines. Annu. Rev. Biochem. 53, 749-790 (1984).
Taylor, F.R. & Cronan, J.E.J. Cyclopropane fatty acid synthase of Escherichia
coli. Stabilization, purification, and interaction with phospholipid vesicles.
166.
Biochemistry 18, 3292-3300 (1979).
Fontecave, M., Atta, M. & Mulliez, E. S-adenosylmethionine:
waste. Trends Biochem. Sci. 29, 243-249 (2004).
167.
DeLano, W.L. The PyMOL Molecular Graphics System.
(lttp :,/ylol
168.
nothing goes to
.so urce forigc.net ).
Chen, D. & Frey, P.A. Identification of lysine 346 as a functionally important
residue for pyridoxal 5'-phosphate binding and catalysis in lysine 2, 3aminomutase from Bacillus subtilis. Biochemistry 40, 596-602 (2001).
- 64 -
Chapter 2
Preliminary crystallographic characterization of biotin
synthase
- 65 -
Abbreviations:
AdoMet
ASU
BioB
DTB
PEG
Tris
S-adenosyl-L-methionine
Asymmetric unit
Biotin synthase
Dethiobiotin
Polyethylene glycol
Tris(hydroxymethyl)aminomethane
- 66 -
The determination of the structure of BioB was a two year-long collaboration with the
laboratory of J.T. Jarrett at the University of Pennsylvania Department of Biochemistry
and Biophysics. Producing good quality crystals of BioB proved to be a challenging
process, requiring approximately 15 months of experimentation to finally obtain 3.4 A
resolution data. Table 2.1 summarizes developments in the crystallization, and the basic
analysis of the structure of BioB. Only one preparation of BioB ever crystallized in my
hands, and I am convinced that something unique about that purification led to
crystallization. The crystallization of BioB by hanging drop vapor diffusion methods was
unusual in several ways which warrant the attention provided by this chapter.
2.1. Protein purification, Fe-S cluster reconstitution, and addition of substrates
Recombinant histidine-tagged Escherichia coli BioB was purified from E. coli and the
Fe-S clusters were anaerobically reconstituted in the Jarrett laboratory as previously
described1 ' 2. Additionally, purified AdoMet was produced in the Jarrett laboratory and
used in the crystallization process. Dethiobiotin (DTB) and all other reagents used in
crystallization were purchased from Sigma. The protein solution contained 20 mg/mL
BioB, 25 mM Tris hydrochloride, pH 8.0, and a four-fold molar excess of AdoMet and of
DTB. Protein solutions and all other crystallization materials were handled in a Coy
anaerobic chamber, under an atmosphere of 5% H2 / 95% Ar.
2.2. Crystallization and preliminary analysis
Crystallization experiments were carried out in a Coy anaerobic chamber at room
temperature,
under an atmosphere of 95% Ar, 5% H 2.
- 67 -
To ensure anaerobiosis, the
chamber contains palladium catalysts mounted on fan systems, which react with H2 gas
and scrub any 02 that may be present in the chamber.
All solutions used in
crystallization were thoroughly purged with Ar before transfer into the anaerobic
chamber. Crystallization trays, pipet tips, and all other supplies and apparatus were
subjected to rounds of vacuum and pressurized Ar in order to exchange any adsorbed
oxygen before being transferred to the anaerobic chamber. PEG solutions, cofactors, and
other light-sensitive reagents were kept in a dark box in the anaerobic chamber.
Table 2.2 shows a list of crystallization screens, commercially available from Hampton
Research Corp., that were used in combination with various crystallization methods in an
attempt to crystallize BioB.
Initial crystallization experiments were done in glass
capillary tubes (Kimble Products KIMAX-51, 1.5 x 90 mm), using the microbatch
method. For these experiments, 3 gL of protein (20 mg/mL BioB, 25 mM Tris pH 8.0)
were mixed with an equal volume of precipitant solution and deposited in a capillary, a 6
!iL reservoir of precipitant solution was deposited opposite the protein/precipitant
mixture, with a small space separating the two liquid plugs, and the capillary was then
sealed with wax. These experiments did not produce protein crystals of observable size.
Due to the difficulties and large amounts of protein that are associated with microbatch
crystallization, this method was soon abandoned in favor of a somewhat unconventional
hanging drop vapor diffusion method described below.
Typically, the hanging drop vapor diffusion method involves a small amount of protein,
which is placed on a siliconized glass cover slip, mixed with precipitant solution, and
- 68 -
sealed over a 0.5-1.0 mL well of precipitant solution. The hanging drop method has the
tremendous advantage of requiring far less protein (1 L or less) per crystallization trial
than the microbatch method, while suffering from the disadvantage of requiring far more
precipitant solution, which must be degassed before introduction to the anaerobic
chamber. The degassing step adds a degree of uncertainty to the experiment, since
degassing likely results in a change in concentration of the solution.
Rather than degassing larger volumes of each component of each screen to use as well
solutions, a solution of 2.5 M (NH4 )2 SO4 was used as a standard well solution, in place of
precipitant solution. The use of ammonium sulfate as well solution was a common
practice, before the availability of commercial crystallization screens. For the purposes
of anaerobic crystallization, this method allows the investigator to reclaim the time that it
would take to degas larger volumes of precipitant solutions, but presents disadvantages as
well.
For instance, small amounts of volatile ammonia may evolve from aqueous
(NH4 )2 SO4 , diffuse into the hanging drop, and change its pH over the course of the
experiment3 .
Likewise, if the precipitant solution contains volatile components, their
concentrations would be expected to fall as diffusion from the hanging drop to the well
solution proceeds. Nevertheless, it was assumed that the salt-equilibrated hanging drop
method would be useful in screening a large number of precipitant solutions. In later
experiments, the potential problem of ammonia volatility was eliminated by using 2.5 M
NaCl instead of a solution of (NH4)2 SO4 .
For a given precipitant solution, wells
containing the NaCl solution generally produced fewer and better quality crystals than
wells containing (NH 4 )2 SO4.
- 69 -
The salt-equilibrated hanging drops were set up as follows: 1 tL of protein (20 mg/mL
BioB, 25 mM Tris pH 8.0) was mixed with 1 tL of precipitant solution (see Table 2.2)
on a siliconized cover slip. The cover slip was placed over a well solution of NaCl or
(NH 4 ) 2 SO4 .
temperature.
Trays were set up and stored in the Coy anaerobic chamber at room
Approximately two weeks into the salt-equilibrated hanging drop
experiments, small yellow-brown crystals were observed to grow in crystallization
condition BioB1 (Table 2.1). Interestingly, BioB forms a very heavy brown precipitate
within an hour of mixing protein with precipitant solution, indicating a relatively rapid
phase transition of the enzyme under these crystallization conditions. The small crystals
appeared to grow from this precipitate. Unfortunately, these crystals were too small to
manipulate, even with a 0.05 mm nylon cryoloop. It was assumed, for reasons discussed
above, that the salt-equilibrated hanging drop method would add an extra layer of
complications to the crystallization experiment, so efforts to reproduce these crystals and
to screen around this condition were undertaken using the conventional method. During
the course of two weeks, no crystals were observed to grow when the well solution was
replaced with precipitant solution. The salt-equilibrated method was then redeployed,
resulting in the successful reproduction of the initial crystals. Screening around the
original condition (by varying the salt concentration and the molecular weight and
concentration of the PEGs used) resulted in a slightly better crystallization condition
BioB2 (Table 2.1). The crystals grown in condition BioB2 were slightly larger and thus
more intensely brown.
As in the original crystallization condition, a heavy brown
precipitate was observed to form several minutes after mixing the precipitant and protein
solutions, and crystals appeared to grow out of this precipitate. Attempts to lessen the
- 70 -
precipitation by decreasing the concentration of PEG resulted in failure to produce
crystals, suggesting once again that precipitate formation is part of the pathway to crystal
growth. The growth of protein crystals out of precipitate is a common phenomenon in
macromolecular crystallography4 .
The crystallization seemed to be extremely sensitive to motion disturbances.
Crystallization trays that were not handled for two weeks after the experiment was started
produced fewer, crystals which were large enough to manipulate with a 0.05 mm
cryoloop. These crystals were used in macro- and micro-seeding experiments. However,
these seeding experiments failed to produce larger crystals.
At this point, crystals grown in condition BioB2 were characterized by SDS-PAGE (Fig.
2.1). Interestingly, lane 3 of Fig. 2.1 contains two distinct bands, one at -39 KDa and a
second band at -36 KDa, instead of one expected single band at -40 KDa (BioB is a
homodimer of -80 kDa).
The -36 KDa band is due to a problem with C-terminal
proteolysis that has now been corrected (J. Jarrett, personal communication). Proteolysis
was occurring at the C-terminal end of the protein just after the last alpha-helix of the
TIM barrel, and although this gel suggests that approximately half of the protein in the
crystals should have the last 31 residues, no electron density is observed for these
residues in either molecule in the structure. It is unclear what the function of this Cterminal region is, since proteolyzed protein shows no reduction in activity compared to
the wild-type enzyme (J. Jarrett, personal communication). It is also not clear whether
proteolysis is an important factor in crystallizing BioB, since the C-terminus of the
- 71 -
crystallographic model does not participate in lattice contacts (Fig. 2.2).
However,
crystallization of C-terminal truncated BioB is probably worth pursuing in order to obtain
higher resolution crystallographic data, since disordered regions of proteins can prevent
the formation of tight lattice contacts.
Crystals produced by condition BioB2 did not diffract synchrotron source X-rays when
cryoprotected and tested for diffraction at 100K. Likewise, no diffraction was observed
when crystals were mounted in capillaries and tested using a rotating copper anode
generator. Attempts were made to improve the crystal quality using detergents and
crystallization additives. Each of Hampton Detergent Screens 1, 2, and 3 was used to
supplement condition BioB2 in crystallization experiments.
Due to the surfactant
properties of detergents, the sitting drop vapor diffusion method was used, in which
protein and precipitant solution are mixed and placed in a small reservoir above the well
solution. Like the successful hanging drop experiments, the sitting drops were also saltequilibrated. Addition of detergents to condition BioB2 had a deleterious effect on the
crystallization, and detergents were generally excluded from further crystallization
experiments. A small molecule additive screen was designed with the hope that additives
might improve the crystallization. The additive screen, whose formulations are shown in
Table 2.3, included salts, polymers, metals, organic solvents, amino acids, sugars,
reductants, cofactors, polyamines, etc. Through screening with additives, it was observed
that substantially larger crystals (Fig. 2.3a ) are grown in condition BioB2 supplemented
with 100 mM glycine and titrated to pH -6.5 (condition BioB3, Table 2.1). As before,
these crystals appeared only after a heavy brown precipitate was observed.
- 72 -
Cryo-cooled crystals grown in condition BioB3 diffracted Cu K, source X-rays, albeit
weakly to -10-11
A resolution
(Fig. 2.4).
Though the resolution is quite low, the
diffraction quality is good, with no evident problems due to cryo-protection or multiple
lattices. Cryo-cooling was accomplished inside the anaerobic chamber by first soaking
the crystal in condition BioB3 supplemented with 20% v/v glycerol for several seconds,
and then plunging the crystal into liquid nitrogen.
50% v/v glycerol was carefully
prepared by degassing for 24 hours in a round-bottom flask with stirring. To ensure that
cryo-protection was not damaging the crystals, a crystal was transferred from the
crystallization drop to a quartz capillary anaerobically, and then tested for diffraction at
room temperature. Crystals tested in this manner produced weak or no discernable
diffraction. A further test was done to see if the resolution limit could be improved by
cryo-cooling crystals in a gaseous stream of nitrogen at 100 K. This method takes
advantage of the transient oxygen stability of BioB in the presence of its substrates5, and
seemed to produce better-quality diffraction than freezing in liquid nitrogen.
Cryo-
cooling in the cold stream was adopted as the standard method of cryo-protecting BioB
crystals.
The best crystals grown in condition BioB3 appeared 1 month after the crystallization
experiment was set up, and required that the crystallization tray not be handled during the
duration of the experiment. Such crystals diffracted synchrotron source X-rays to afford
data to 5.5 A resolution (Table 2.4). Data were processed with the programs DENZO and
SCALEPACK6 . While these data were not of high enough resolution to allow us to solve
the structure, the data were useful for the purposes of beginning the characterization of
- 73 -
the crystals. Crystals are trigonal (P321, P3 121, or P3 221), with a=b=155.3 A, c= 91.3 A.
Unfortunately, no data were collected along (001), and it was thus impossible to properly
identify a 31 or 3 2screw axis at this time. Given the dimensions of the unit cell and the
molecular weight of BioB, a solvent content of -53% (corresponding to a Matthews
coefficient7 of Vm=2.65A3 /Da and 3 molecules per ASU) or -69% (corresponding to
Vm=3.98 A 3/Da and 2 molecules per ASU) was expected.
Since BioB is a homodimer
and the crystals diffract to low resolution, we suspected that the solvent content was
indeed 69%, and this turned out to be correct. Data collected at the Fe absorption peak
wavelength was processed with DENZO and SCALEPACK6 , and SOLVE8 was used to
search for Fe sites in the possible space groups. In (P3121, P3221), four Fe sites were
identified (Fig. 2.6), consistent with expected set of two 4Fe-4S and two 2Fe-2S clusters
per BioB dimer2. Spacegroup P3121 and the sites that were located turned out to be
correct (see below), illustrating the value of low resolution (5.5 A) data in preliminary
crystal characterization.
To improve upon the 5.5
A resolution
limit, the crystallization of BioB was revisited.
Upon re-examination of the conventional hanging-drop crystallization experiments, small,
irregularly shaped crystals were observed to grow in condition BioB4. Based on previous
results, it was inferred that the addition of glycine to the precipitant and the use of the
salt-equilibrated hanging drop method might produce better crystals. Condition BioB4
was optimized to give condition BioB5, consisting of 0.1 M Tris hydrochloride, 0.1 M
glycine, 0.2 M MgC12, 20% w/v polyethylene glycol 1000, which was titrated to a final
pH of -6.5.
Once again, trays were set up in a Coy anaerobic chamber at room
- 74-
temperature, with 1 AtLof protein (20 mg/mL BioB, 25 mM Tris pH 8.0) added to 1 L
of precipitant solution on a siliconized cover slip, and equilibrated over a well containing
0.5 mL 2.5 M NaCl. Crystals grown in condition BioB5 (Fig. 2.3b) were the largest
BioB crystals obtained to date and diffracted synchrotron
source X-rays to 3.4 A
resolution (Table 2.5). These crystals belong to space group P3121 (six ASUs), with 2
molecules per ASU (Z=12) and a = b =155.69 A, c = 90.88 A, V=1,908,412 A3 .
Systematic
absences
distinguished
(P3 121,
P3 221)
from
p321,
and
P3 121
was
differentiated from P3221 by inspection of a helices.
As with previous crystallization conditions, extensive macro- and micro- seeding
experiments under condition BioB5 failed to produce larger or better-quality crystals, and
crystals always appeared several days to weeks after the formation of a heavy brown
precipitate in the crystallization drop. Using Tris that was buffered at pH 7.0 (instead of
titrating the final precipitant solution to pH -6.5) resulted in smaller crystals that
diffracted relatively poorly, suggesting that Tris is serving as an additive rather than as an
effective buffer in condition BioB5. Data collected from crystals grown in condition
BioB5 allowed us to solve the structure of BioB. The protein's intrinsic Fe cofactors
were used to phase the structure by multiwavelength anomalous dispersion (MAD)
methods, as described in chapter 3.
- 75 -
2.3. Tables
Table 2.1. Timeline of BioB crystallization.
Date
Development
Crystallization
condition
Precipitant formulation
Well solution
(0.5 mL, 2.5 M)
Jul.
2001
Small crystals
BioB1
0.2 M KC1
0.01 M MgC12
0.05 M Na-Cacodylate
(NH 4 )2 SO4
pH 6.5, 10% w/v PEG
4000
Sep.
Crystals large enough
2001
to manipulate
BioB2
0.2 M MgC12
NaCl
0.1 M Na-Cacodylate
pH 6.5 20% w/v PEG
1000
Nov.
2001
-11 A diffraction (Cu
Kradiation)
BioB3
0.2 M MgC12
0.1 M Na-Cacodylate
pH 6.5
NaCl
0.1 M glycine
20% w/v PEG 1000
5.5 A data
(synchrotron
radiation), location of
two putative Fe-S
cluster sites per BioB
monomer
BioB3
Crystals observed in
previous
conventional hanging
drop experiments
BioB4
Aug.
3.7 A data
BioB5
2002
(synchrotron
radiation)
Jan.
2002
NaCl
0.2 M MgC12
0.1 M Tris-HCI pH 7.0
NaCl
12% w/v PEG 3000
0.1 M Tris-HCl
0.1 M glycine
0.2 M MgC12
20% w/v PEG 1000
NaCl
final pH -6.5
Sep.
2002
3.4 A data
(synchrotron
radiation)
NaCl
BioB5
- 76 -
Table 2.2. Hampton Research screens used to find BioB crystallization conditions.
Screen
Microbatch
capillaries
Conventional
hanging drop
Tried
(NH 4 ) 2S0 4
-equilibrated
hanging drop
NaClequilibrated
hanging drop
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Tried
Natrix
Tried
Tried
Tried
Tried
MembFac
Tried
Tried
Tried
Tried
Not Tried
Tried
Not Tried
Not Tried
Not Tried
Tried
Tried
Tried
Not Tried
Tried
Tried
Tried
Not Tried
Tried
Tried
Tried
Not Tried
Tried
Not Tried
Not Tried
Sparse-matrix
Crystal
Screen
Crystal
Screen 2
Crystal
Screen Cryo
Crystal
Screen Lite
PEG/Ion
screen
Gridscreens
Grid Screen
(NH 4 )2 SO4
Grid Screen
PEG 6000
Grid Screen
MPD
Grid Screen
PEG/LiCl
Grid Screen
NaCl
For each screen listed, one or more crystallization methods were used in an attempt to
crystallize BioB. For crystallization drops that were equilibrated against either
(NH4 )2 SO4 or NaCi wells, the well solution was 2.5 M.
- 77 -
Table 2.3. Additives used to optimize BioB crystallization conditions.
#
Additive
Stock
Unit
Conc.
Conc.
#
M
Additive
Stock
Conc.
Unit
Conc.
41
42
43
NaFormate
NaH 2PO 4
D-Sucrose
1.0
1.0
M
M
30
% w/v
% v/v
% v/v
% v/v
% v/v
44
NaKTartarate*4H 20
1.0
M
45
D-Sorbitol
46
Dioxane
30
30
% w/v
% v/v
47
Na 2 EDTA
48
49
CsCl
KSCN
50
51
52
NaOAc
NaCitrate
BaC12
0.1
1.0
1.0
1.0
1.0
0.1
M
M
M
M
M
M
53
54
55
56
PEG 3000
PEG 5000 MME
PEG 20,000
Jeffamine M 600 pH 7.0
30
30
30
50
% w/v
% w/v
% w/v
% v/v
57
58
L-Lysine
L-Arginine
1.0
1.0
M
M
59
60
61
Dextran Sulfate
Decanedioic acid
6-Aminocaproic Acid
0.15
% w/v
Sat'd soln.
62
63
64
1,6-Diaminohexane
1,8-Diaminooctane
1,5-Diaminopentane
65
Azelaic acid
66
67
2
MnC12
3
ZnC12
4
Ethylene Glycol
0.1
0.1
0.01
30
5
Glycerol
30
6
7
MPD
PEG 400
30
8
9
10
11
12
13
14
15
16
17
18
Urea
Ethanol
Isopropanol
Methanol
(NH4 )2 SO4
KCI
LiCI
NaCl
CH 3CN
Acetone
DTT
0.1
30
30
30
1.0
1.0
1.0
2.0
M
40
40
% v/v
% v/v
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.245
M
M
M
% v/v
M
M
M
M
M
M
M
68
(Glycyl) 2-glycine
(Glycyl) 2-glycine
(Glycyl) 3-glycine
1,10-Diaminodecane
69
70
Palmitic acid
L-Methionine
71
72
73
L-Serine
L-Threonine
L-Glutamate
1
1
1
1
M
M
M
M
1
CaC1 2
50
M
M
% v/v
% v/v
% v/v
M
M
M
M
19
CdC12
20
CoC12
21
22
23
Saturated
SrC12
NiC12
24
25
CuSo 4
Spermine
26
27
Spermidine.3HCl
Octyl Glucoside (n-Octyl-P-DGlucoside)
28
LDAO (Lauryldimethylamine
0.02
29
30
Oxide)
Dimethylsulfoxide
Glutathione (reduced)
30
% v/v
M
31
32
33
Glycine
Mg(OAc) 2
D-Glucose
0.025
1.0
0.1
34
Guanidine.HCl
35
Hexaaminecobalt(III)Trichlori
de
Na 2Malonate
FeC12
10
M
M
20
1.0
1.0
M
M
Sat'd soln.
0.5
M
Sat'd soln.
0.3
M
0.3
M
0.1
M
Sat'd soln.
Sat'd soln.
w/v
36
1.0
0.1
M
M
74
L-Glutamine
1
M
75
L-Aspartate
1
M
1.0
M
76
Hexadecyltrimethylammoni
0.01
M
0.01
0.01
0.01
M
M
M
um bromide
37
38
Li 2SO4
KI
39
40
Beta alanine
NaBr
1.0
1.0
0.1
M
M
M
1.0
M
- 78 -
77
78
Adenine
Biotin
79
Pyridoxal 5'-phosphate
Table 2.4. Data collected from a single crystal grown in condition BioB3.
Remote
Fe Peak
Wavelength (A)
1.2
1.736
Exposure time (s)
Oscillation range (deg)
60
1
80
1
Resolution (A)
30-5.5
30-5.5
RSym' 2 (%)
6.7 (30.3)
8 (41.6)
Unique reflections
40253
7584
Redundancy
6.3
3.2
I/cF(I) 2
11.2 (5.14)
95.0 (95.5)
7.5 (2.27)
95.8 (93.8)
Completeness (%)2
Data were collected at Beamline 9-2, Stanford Synchrotron Radiation Laboratory, at 100
- <I(hkl)>] /k<I(hkl)>, where Ii(hkl) is the ith measured
K. 'Rsy = YkI[,i(hkl)
diffraction intensity (no a cutoff) and <I(hkl)> is the mean intensity of the miller index
(hkl). 2 Values for the highest resolution bins are in parentheses. 3For this wavelength,
Friedel pairs were merged during data processing.
- 79 -
Table 2.5. Data collected from crystals grown in condition BioB5 and used to solve the
structure of BioB.
Wavelength (A)
Resolution
Unique
range (A) reflections
Redundancy
Completeness (%)
(
R
12
s
1.30000
100- 3.7
26,018
3.4
99.5 (100.0)
11.8 (2.3)
6.7 (37.5)
1.73827
100- 4.1
19,002
3.1
98.3 (99.7)
14.6 (1.9)
6.4 (29.7)
1.74150
100 - 4.5
14,030
2.8
97.1 (95.7)
14.0 (2.6)
8.2 (36.7)
1.100003
100 - 3.4
17,465
5.4
98.1 (87.9)
15.8 (3.8)
6.6 (25.9)
Data were collected at 100 K, using an inverse beam geometry, at the Advanced Light
Source beamline 5.0.2 (1.30000A,
1.73827A, and 1.74150A wavelengths)
and the
National Synchrotron Light Source beamline X25 (1.10000 wavelength), both equipped
with a charge-coupled device detector (Area Detector Systems Corp). Data were
processed and scaled with DENZO 6 , SCALEPACK 6 , and XDS9 .
Values for the highest
resolution bins are in parentheses. 2 Rsy = DEkl[li(hkl) - <I(hkl)>]l /Ew<I(hkl)>,
where Ii(hkl) is the ith measured diffraction intensity (no a cutoff) and <I(hkl)> is the
mean intensity of the miller index (hkl). 3 For this wavelength, Friedel pairs were merged
during data processing.
- 80-
2.4. Figures
1
2
3
4
B
40
50
37
25
o
4
Figure 2.1. SDS-PAGE analysis of BioB crystals grown in condition BioB2. A sample
of the crystallization drop (lane 2) was run on a 15% polyacrylamide gel alongside BioB
crystals that were first washed with fresh precipitant solution and then dissolved in
distilled water (lane 3). 50, 37, and 25 kDa molecular weight markers (lanes 1 and 4) are
labeled.
- 81 -
Figure 2.2. Lattice contacts in the final structure of BioB, showing the last observable
residue in both molecules in the ASU (each molecule is depicted as an alpha carbon
trace). The model is truncated at residue 315 since no electron density was observed for
the last 31 residues of the protein. Symmetry related molecules are show in shades of
pink, purple, and red. The C-termini do not participate in lattice contacts.
- 82 -
(a)
(b]
-O.a mm
Figure 2.3. BioB crystals. (a) Crystals grown in condition BioB3 were transferred to a
capillary to check for diffraction at room temperature. The crystals measured less than
10 microns in the largest dimension. The heavy brown precipitate, characteristic of all
BioB crystallizations, is prominent in this image. (b) Crystals grown in condition BioB5.
Crystals appeared as bloated hexagonal rods with a maximal radius of -50 microns and a
maximal length of -200 microns. Such crystals produced the data used to solve the
structure.
- 83 -
Figure 2.4. First diffraction from a BioB crystal. (a) Crystals grown in condition BioB3
were cryoprotected by briefly soaking in precipitant solution supplemented with 20% v/v
glycerol, followed by plunging into liquid nitrogen in the anaerobic chamber. The
diffraction limit is -10-11 A when tested with Cu Ks radiation. (b) zoomed-out view of
the image in part a.
- 84-
Figure 2.5. 5.5 A data from crystals grown in condition BioB3. Data were collected at
SSRL beamline 9-2. Though the resolution is low, the diffraction quality is generally
good.
- 85 -
(a)
(b)
Fe-S 2
4
A
~
_ *,' A
IU
Fe-S
4
I,
1
g
k
,-%
~~~~~~~~~~~~~-1
i e, S1
Fe-S
3
Figure 2.6. Location of putative Fe sites in BioB. (a) Data collected from crystals
grown in condition BioB3 at the Fe absorption peak wavelength were used to locate
putative Fe-S clusters in unit cell. The unit cell is visualized parallel to the c axis. This
number and arrangement of Fe-S sites is consistent with a structure that reasonably fills
the unit cell and has sensible lattice contacts. (b) Diagram showing the cluster sites in the
asymmetric unit, showing intercluster distances.
- 86 -
2.5. References
1.
Ugulava, N.B., Gibney, B.R. & Jarrett, J.T. Iron-Sulfur cluster Interconversion in
Biotin Synthase: Dissociation and Reassociation of Iron during Conversion of
[2Fe-2S] to [4Fe-4S] Clusters. Biochemistry 39, 5206-5214 (2000).
2.
3.
4.
5.
6.
7.
Ugulava, N.B., Gibney, B.R. & Jarrett, J.T. Biotin synthase contains two distinct
iron-sulfur cluster binding sites: chemical and spectroelectrochemical analysis of
iron-sulfur interconversions. Biochemistry 40, 8343-8351 (2001).
Seinfeld, J.H. & Pandis, S.N. Atmospheric Chemistry and Physics, 1326 (John
Wiley and Sons, New York, 1998).
http://xray.bmc.uu.se/-terese/crystallization/tutorials/tutorial2.html.
Ugulava, N.B., Frederick, K.K. & Jarrett, J.T. Control of AdenosylmethionineDependent Radical Generation in Biotin Synthase: A Kinetic and Thermodynamic
Analysis of Substrate Binding to Active and Inactive Forms of BioB.
Biochemistry 42, 2708-2719 (2003).
Otwinowski, Z. & Minor, W. Processing of X-ray diffraction data collected in
oscillation mode. Methods Enzymol. 276, 307-326 (1997).
Matthews, B.W. Solvent content of protein crystals. J. Mol. Biol. 33, 491-497
(1968).
8.
Terwilliger, T.C. & Berendzen, J. Automated structure solution for MIR and
MAD. Acta. Cryst. D55, 849-861 (1999).
9.
Kabsch, W. Automatic processing of rotation diffraction data from crystals of
initially unknown symmetry and cell constants. J. Appl. Cryst. 26, 795-800
(1993).
- 87 -
Chapter 3
Crystal structure of biotin synthase
- 88 -
Abbreviations:
2,3-LAM
Ado.
AdoCbl
AdoH
AdoMet
ASU
BioB
DTB
FNR
HemN
MAD
NADPH
NCS
PFL
PLP
Lysine 2,3-aminomutase
5'-deoxyadenosyl radical
Adenosylcobalamin, coenzyme B12
5'-deoxyadenosine
S-adenosyl-L-methionine
Asymmetric unit
Biotin synthase
Dethiobiotin
Flavodoxin/NADPH oxidoreductase
Coproporphyrinogen III oxidase
Multiwavelength anomalous dispersion
Reduced nicotinamide adenine dinucleotide phosphate
Non-crystallographic symmetry
Pyruvate formate-lyase
Pyridoxal 5'-phosphate
SAM
S-adenosyl-L-methionine
TIM
Tris
Triosephosphate isomerase
Tris(hydroxyethyl)aminoethane
- 89 -
3.1. Introduction
Biotin synthase (BioB) catalyzes the final step in the biotin (vitamin B8 ) biosynthetic
pathway, the conversion of dethiobiotin (DTB) to biotin. This remarkable reaction
requires the insertion of a sulfur atom between non-activated carbons C6 and C9 of DTB
(Fig. 3.1). BioB contains unusual iron-sulfur clusters that play a role in sulfur insertion
through an AdoMet-dependent radical-based mechanism.
Like all AdoMet radical
enzymes, BioB contains a conserved CxxxCxxC sequence motif that coordinates an
essential redox-active 4Fe-4S cluster, with AdoMet coordinating a unique Fe of this
cluster' -5. The crystal structure of BioB in complex with AdoMet and DTB is presented
in this chapter. This work begins to address the disputed cofactor content of BioB, the
two hypothetical mechanisms for the BioB reaction, and AdoMet binding by the AdoMet
radical enzymes. The structure also suggests an evolutionary relationship between the
AdoMet- and AdoCbl-dependent radical enzymes, which is discussed in further detail in
chapter 5. A modified version of this chapter was published in January 20046.
In the reaction catalyzed by BioB, there is general agreement that Ados generated from
AdoMet oxidizes DTB, as demonstrated by Marquet and coworkers7 , but the number and
types of Fe-S clusters and other cofactors involved in the reaction have been a subject of
controversy - 15.
Protein preparation-dependent cofactor differences have led to two
mechanistic proposals for the method of sulfur insertion in BioB. The "hybrid cluster
model", put forth by Jarrett and coworkers16,involves the use of a 2Fe-2S cluster as the
immediate sulfur donor for biotin (Fig. 3.2a), and is consistent with Marquet's pioneering
- 90 -
34S
isotopic labeling studies17 and with the observed destruction of a 2Fe-2S cluster that
accompanies BioB turnover 6 ' 18. The hybrid-cluster model of Jarrett and coworkers lead
to a proposal which requires two molecules of AdoMet and two reducing equivalents
(Fig. 3.2a). The stoichiometry of AdoMet and the sulfur source in this mechanism are
both consistent with initial studies from the Marquet laboratory 7 ' 19, but conflict with
studies from Fontecave and coworkers 1,20 . The proposal of Fontecave and coworkers is
based on the activity of BioB from which the 2Fe-2S cluster has been removed, and
suggests that S is provided from an enzyme-bound persulfide, generated via an intrinsic
pyridoxal-5'-phosphate (PLP) -dependent cysteine desulfurase activity"' 20 (Fig. 3.2b).
Recently, Johnson and coworkers have obtained evidence that BioB does not bind PLP21,
casting doubt on the intrinsic cysteine desulfurase activity proposed by Fontecave's
group l20. Spectroscopic data from the Johnson and Huynh groups support the presence
of two distinct Fe-S clusters in BioB, a 4Fe-4S cluster and a 2Fe-2S clusterl 4' 62 22
These studies are consistent with sulfur donation from the 2Fe-2S cluster to DTB and
destruction of the 2Fe-2S cluster (the hybrid cluster model), but, in contrast to Jarrett and
coworkers, the destruction of the 2Fe-2S cluster was found to be faster than the formation
of biotin22. The mechanism proposed by Jarrett and coworkers can still account for the
results from the laboratories of Johnson and Huynh, but the kinetic analysis does not
exclude additional steps between the 2Fe-2S cluster destruction and formation of biotin.
Our results, presented below, also favor a mechanism that invokes a role for a 2Fe-2S
cluster over a mechanism that involves PLP. Irrespective of whether BioB preparations
contain a 2Fe-2S cluster or PLP, assay mixtures produce one or fewer turnovers per
monomer, over the course of several minutes to hours, although BioB from Arabidopsis
- 91 -
is reported to catalyze >2 turnovers per hour (Table 3.1). Due to low enzymatic activity,
it has been suggested that BioB is either subject to strong product inhibition by AdoH20 ,
or that it may be a "suicide enzyme," i.e. a stoichiometric reactant rather than a true
catalyst.
- 92 -
3.2. Materials and methods
The crystallization of BioB was discussed in detail in chapter 2. This section describes
crystallographic methods that were used to arrive at the final model of BioB.
The experimental electron density map was calculated using the phase information
derived from four Fe sites in the asymmetric unit (ASU), corresponding to one 4Fe-4S
clusters and one 2Fe-2S clusters for each of the two molecules in the ASU. The program
SOLVE2 3 was used to refine the sites and to calculate the experimental electron density
map. The non-crystallographic symmetry (NCS) operator that relates thee two molecules
in the ASU was found using the program LSQKAB2 4 , after manually placing a helices
into the density corresponding to both molecules in the ASU. The experimental electron
density map, calculated to 3.7 A resolution, was subjected to solvent flattening and
twofold NCS averaging using DM25 , which resulted in an electron density map of highenough quality to distinguish the 2Fe-2S cluster from the 4Fe-4S cluster of the BioB
monomer, and to model the clusters in their respective densities (Fig. 3.3).
Using
SOLVE23 , the coordinates of the twelve Fe atoms in the ASU (4 Fe for each of two 4Fe4S clusters + 2 Fe for each of two 2Fe-2S clusters) were refined and used to calculate
phases. The electron density map that resulted was subjected to solvent flattening and
twofold NCS averaging, giving the map that was used to build the initial model (Fig.
3.4). After several rounds of iterative model building using the program 026, refinement
of the model with CNS2 7 , and rebuilding of the model into sigma-A weighted, phasecombined electron density maps, the resolution of the experimental map was extended to
3.4 A (using the programs SIGMAA, SFALL, and FFT from the CCP4 package28 ), and
- 93 -
solvent flattening and twofold NCS averaging were performed again. Toward the end of
the model building process, electron density maps with coefficients of 2IFol-IFcIand IFolIFc and phases calculated from the model were used, along with the phase-extended
experimental electron density map.
The model was refined against 3.4 A data using the program CNS27 , to a final R = 25.6%
and Rfree= 30.0% on F (Table 3.2). Early in the refinement, strict NCS constraints were
applied to the two molecules in the ASU; these were relaxed to moderate restraints (with
a weight of 400) during the last few rounds of refinement. The refined structure of the
dimeric enzyme contains residues 4-315 (molecule A) and 3-315 (molecule B), two DTB,
two AdoMet, two Fe4 S 4 clusters, two Fe2S 2 clusters, and one Tris molecule. There is no
electron density for the N-terminal histidine tag, the first 2 or 3 residues at the Nterminus, and the last 31 residues at the C-terminus in either molecule.
The final model has 99.3% of all residues residing in the allowed region of the
Ramachandran plot (Fig. 3.5), with 79.8% (molecule A) and 79.8% (molecule B) in the
most favored region, 18.3% (molecule A) and 17.9% (molecule B) in the additional
allowed regions, 1.5% (molecule A) and 1.5% (molecule B) in generously allowed
regions, and 0.7% (molecule A) and 0.5% (molecule B) in disallowed regions, as
calculated by PROCHECK2 9 . One of the residues in the disallowed regions, Asn311, is
at the conformationally flexible C-terminus. The other residue, Asp 155, is in position to
interact with the ribose moiety of AdoMet, and therefore its unusual geometry is likely to
- 94 -
be functionally significant. The equivalent residue in HemN, another AdoMet radical
enzyme, also has disallowed Ramachandran angles30 .
3.3. Results
3.3.1. Overallstructureand the locationof Fe-S clusters
The fold of each subunit of the BioB homodimer is a TIM barrel, with two additional
helices at the N-terminus and a disordered region at the C-terminus (Figs. 3.6a, 3.7a and
b). The arrangement of the two TIM barrels allows access to the bottom and to the top of
both subunits, forming an elongated structure (Figs. 3.6a and 3.8). BioB joins at least 26
other protein superfamilies incorporating TIM barrel architecture3 1 and four AdoCbldependent enzymes that use this fold for catalysis of radical-based chemistry (see chapter
1). However, to the best of our knowledge, this is the first example of Fe-S clusters
bound within and on top of a TIM barrel.
A 4Fe-4S cluster is located at the C-terminal end of the TIM barrel, far (-30 A) from the
dimer interface, and a 2Fe-2S cluster is located deep inside of the barrel, -25
A from
the
C-terminal end (Fig. 3.6a, 3.7a and b). The active site, containing AdoMet and DTB, is
situated between these two clusters (Figs. 3.6a and b, Fig. 3.7a and b). This Fe-S cluster
arrangement suggests an explanation for the reported differences in Fe-S cluster content
of BioB; a surface Fe-S cluster may be more readily lost and reconstituted than a deeply
buried cluster. The 2Fe-2S cluster is observed by M6ssbauer spectroscopy8' 32 in whole
cells containing
protein9
2' 3'1 5,33.
recombinant
BioB, and is always retained
in the initial purified
Treatment of BioB with strong chemical reductants and metal
chelators, followed by Fe reconstitution, yields a 2Fe-2S-depleted form of BioB, with an
- 95 -
intact 4Fe-4S cluster9 0,
'1 '4 34 '3 5. Reconstituted protein not subjected to a reduction or
chelation step (such as used in this study) contains both a 4Fe-4S and a 2Fe-2S
cluster 8 1' 4 1' 5' 2 1.
The lattice packing of BioB crystals is shown in Fig. 3.9. Each BioB subunit contacts
three neighboring subunits of neighboring ASUs. The interaction between subunits of
neighboring ASUs is sparse and does not suggest a physiological docking surface for
other proteins, such as flavodoxin.
3.3.2. A PLP binding site is not observed in the structure
With respect to the proposal that PLP is a cofactor in BioB" 2 0 , we did not add PLP
during protein purification or crystallization and do not find it in our structure.
Additionally, the structure does not reveal an obvious binding site for PLP. Indeed,
cysteine desulfurases typically bind PLP via an imine linkage to a Lys residue, and all
Lys residues, including the one highly conserved Lys (K49 in BioB) are found at the
surface of the protein (Fig. 3.10), far from the putative site of a cysteine persulfide" l and
the observed DTB binding site. Furthermore, no obvious phosphate-binding site can be
found in the structure. In accord with studies from the Jarrett ' 6 , Marquet1 7' 3 6 , and
Johnson
and Huynh21 ' 22 groups,
coworkers''
20 ,
and in contrast
to results
from Fontecave
and
this X-ray structure analysis favors a mechanism that invokes a role for an
2Fe-2S cluster over one that requires PLP.
- 96 -
3.3.3. The novel 2Fe-2S cluster of BioB
The 2Fe-2S cluster of BioB is unique both in terms of its coordination by
-strand
residues in the core of a TIM barrel (Figs. 3.6a and b, 3.7a and b, 3.10), as well as in
terms of the identity of the amino acid ligands: C97, C128, C188, and R260, (Figs. 3.6b,
3.11a-c, 3.12a-g). All four of these residues are conserved, and the assignment of the
fourth ligand as Arg is unambiguous, even at the moderate resolution of this structure; the
amino acid sequence of the TIM barrel has been fully assigned with no breaks or gaps,
and the shape of the omit electron density at position 260 is completely consistent with an
Arg sidechain (Fig. 3.1 b).
Our ligand assignment is consistent with Johnson and
coworkers' spectroscopic observation of incomplete cysteinyl ligation for the 2Fe-2S
cluster9'2 1 . The assignment of an Arg ligand to a metal is unprecedented in biology,
although guanidine and guanidine-like species have been observed to ligate metals in
small molecules (Fig. 3.13a-e and Table 3.3; discussed below).
3.3.4. The 4Fe-4S cluster and AdoMet
Proteins are classified as members of the AdoMet radical superfamily based on the
presence of the sequence motif CxxxCxxC (C53, C57, and C60 for BioB) 1. This motif is
contained in a 28-residue loop extending from P-strand 1 to helix 1 of the TIM barrel,
and the three cysteines coordinate three of the four irons of the 4Fe-4S cluster, as
expected from mutagenesis studies33 3 7 (Fig. 3.11a). AdoMet is the fourth ligand to the
cluster and binds as an N/O chelate to a unique Fe position, through the AdoMet amino
group nitrogen and carboxyl group oxygen (Fig. 3.1 la, Table 3.4).
- 97 -
This binding mode is consistent with spectroscopic results for PFL activase2'3 , and is
likely to be common to AdoMet radical proteins, since spectral changes associated with
AdoMet binding have been reported for several family members3 8 40.
In a report
published at the same time as the BioB structure, Schubert and coworkers noted the same
mode of AdoMet binding in the crystal structure of HemN30 , another AdoMet radical
' , the
enzyme. Also consistent with spectroscopic data on BioB and PFL activase3 41
sulfonium of AdoMet does not appear to be a ligand to Fe and is -4.0 A away from the
nearest cluster Fe.
Interestingly, BioB is the only known AdoMet-binding TIM barrel31 ' 42. AdoMet is bound
to BioB in an extended conformation, stretching across the top of the barrel, such that
contacts are made to AdoMet by residues in or following O-strands 1, 2, and 4-6 (Fig.
3.6b, Table 3.5). The net result of these interactions is that AdoMet is completely buried
from solvent and appears to be ideally positioned for both electron transfer from the 4Fe4S cluster and for hydrogen atom abstraction from DTB (discussed below).
Some contacts to AdoMet are made by residues that are homologous across several
AdoMet radical enzymes. For example, 1192, which makes a van der Waal contact to
AdoMet (Fig. 3.1 la), is part of a short sequence motif (192-IxGxxE-196) that is located
near the AdoMet binding site. This sequence corresponds to MxGxxE in LipA, VxGxxD
in PFL activase, LxGxxD in 2,3-LAM, etc. (Fig. 3.14).
Additionally,
almost all
sequenced AdoMet radical enzymes have an aromatic residue at the penultimate position
- 98 -
of the CxxxCxxC sequence. In BioB, this is residue Y59, which forms re-stacking
interactions with the adenine ring of AdoMet. Other contacts to AdoMet are made by
D155, which hydrogen bonds to the 2' and 3' -OH groups of the ribose ring, R173, which
forms a salt bridge with the AdoMet carboxylate, and N153, which hydrogen bonds to
both AdoMet and DTB.
3.3.5. The methionyl moiety
In our structure, the adenine and ribose moieties of AdoMet fit the electron density well,
whereas the broad electron density of the methionyl moiety indicates some ambiguity in
the exact location of the sulfonium (Fig. 3.11a). The electron density may suggest that
we have a combination of cleaved and uncleaved AdoMet, presumably resulting from
radiolytic reduction due to long exposures of the crystal to synchrotron radiation.
Alternatively, the breadth of the methionine might be explained by racemization at the
sulfonium of AdoMet, which is well documented4 3 . A racemic mixture of (S,S)-AdoMet
(which is not found in biological systems4 4 ) and (R,S)-AdoMet
structure of HemN30 .
is observed in the crystal
A more detailed analysis of the methionyl moiety awaits the
availability of higher resolution data.
3.3.6. The dethiobiotinbindingsite
The enzyme preparation used for crystallization has DTB bound4 5 , and we observe
electron density in the active site that fits this substrate (Fig. 3.11c). DTB binds in the
core of the TIM barrel, between the 2Fe-2S cluster and AdoMet. In contrast to the
reported stoichiometry of bound substrates4 5, we find one DTB and one AdoMet in each
- 99-
BioB subunit. This discrepancy is currently under investigation. DTB binding to BioB is
requires the presence of AdoMet4 5 , and we find that DTB makes substantial van der Waal
contacts with AdoMet, covering 50% of its surface. These interactions include the
stacking of the carboxylate tail against the AdoMet adenine and the stacking of the DTB
ureido ring with the AdoMet ribose. DTB makes a series of contacts with protein as well
(Table 3.6), the most important of which is the bidentate interaction with N222 that may
play a role in orienting DTB for hydrogen atom abstraction.
Finally, the DTB
carboxylate interacts with the backbone amides of T292 and T293, as well as the side
chain of T292 (Fig. 3.15). Interactions with these residues could serve to close a loop
over the top of the barrel upon DTB binding, helping to seal the barrel for the radicalmediated reaction that follows.
-100-
3.4. Discussion
3.4.1.Arginine as an Fe ligand
The assignment of an Arg ligand to a metal is unprecedented in biology, although
guanidine species have been observed or proposed to ligate Co"', Os"', Pt 1 , Nil", and Zn"
in small molecules in aqueous solution at pH -3.5 - 104649 (Fig. 3.13, Table 3.3). In
BioB, the metal-nitrogen distance (-2.3-2.4 A) is longer than those observed in the small
molecule crystal structures (1.94 - 2.07 A, Table 3.3). However, at our current resolution
limit of 3.4 A, any discussion of Fe-N distance is premature. It has been predicted that
Arg could serve as a metal ligand in a protein environment where the Arg sidechain were
uncharged4 8, and in BioB, R260 has an unusual environment that should indeed alter its
pKa. It is buried in the center of a TIM barrel and has a high number of potential
hydrogen-bonding partners (S43, S218, S283, and R95, Fig. 3.1 lb). The conservation of
Arg as an Fe-S cluster ligand in BioB, rather than the more common Cys or His residues,
is peculiar and suggests an important role for this residue in modulating the properties of
the cluster or in facilitating catalysis. The choice of metal ligands may be important for
preserving charge neutrality within the core of the (a/) 8 barrel, since one common
mechanism of Fe-S cluster redox modulation by backbone-NH-to-S hydrogen bonds is
not possible with this structural motif. All surrounding main-chain amides are involved
in hydrogen bonding between the -strands of the barrel and can not interact with the
cluster. Thus, the ligands and neighboring sidechains are likely to play a significant role
in redox modulation, and an Arg (or His) ligand would be a better choice than Cys in
terms of reducing the overall net negative charge of this buried cluster. The use of Arg,
- 101 -
rather than His, is more difficult to explain. In part, the answer may be structural, in that
the long length of the Arg side-chain allows for a barrel that is less compact and able to
fit both substrates. However, R260 may be conserved as a ligand to the 2Fe-2S cluster
for catalytic rather than structural reasons. During turnover of BioB, the 2Fe-2S cluster is
1 6'17'36 . In our structure, the R260 side-chain
destroyed as one S is transferred into biotin
could rearrange to bridge the two Fe atoms and facilitate the proposed S transfer. Our
knowledge of how Fe-S clusters are assembled in biological systems is still in its infancy,
and it will be interesting to discover if any of the features observed in the BioB structure
are mimicked in proteins involved in cluster assembly.
3.4.2. Structural insight into the reaction catalyzed by BioB
Although the elucidation of a detailed enzyme mechanism awaits further biochemical
studies, the structure does provide insight into key steps of the BioB reaction. The
observed cooperativity of substrate binding45 can be explained by the extensive
interaction between AdoMet and DTB (described in section 3.3.6), and by the interaction
of both substrates with N153. With respect to radical generation, the electron transfer
from flavodoxin to the 4Fe-4S cluster is made possible by the binding of the cluster close
to the protein surface (-6-7 A). The subsequent electron transfer from the 4Fe-4S cluster
to the AdoMet sulfonium is facilitated by direct O/N coordination of the AdoMet to Fe,
thereby restraining the position of the sulfonium to the proximity of the 4Fe-4S cluster.
The resulting formation of Ado- is likely coupled to hydrogen atom abstraction from C9
of DTB, and we find that DTB is positioned accordingly, with C9 -3.9 A away from the
5' carbon of AdoMet. Closure of the biotin thiophane ring requires the abstraction of a
- 102-
second hydrogen atom from C6 of DTB. Deuterium transfer from of the C6 position of
(2 H)DTB has demonstrated that a second Ado. is likely also responsible for this second
hydrogen atom abstraction7, although the requirement for two equivalents of AdoMet has
recently been disputed2 0 . In our structure, C6 of DTB is positioned -4.1
A from the 5'
carbon position of AdoMet, suggesting that Ado. also accomplishes the second hydrogen
atom abstraction. Exactly how structural changes of the enzyme facilitate the release of
AdoH and Met and the binding of a second AdoMet molecule remains to be determined.
The rearrangements involved in this process will likely involve the movement of loops at
the C-terminal end of the TIM barrel.
The sulfur insertion step is the most controversial part of the mechanism of BioB. The
structure analysis shows that the 2Fe-2S cluster is ideally positioned to play a role in
sulfur insertion. The closest bridging S of the 2Fe-2S cluster is only 4.6 A away from C9
of DTB, a position which is consistent with transfer of this sulfur to biotin as suggested
by results from
3 4 S-labeling
experiments17 and the observation of 2Fe-2S cluster
degradation during turnover6 ' 18 ' 22
3.4.3. Questions arisingfrom the structure of BioB
If the in vivo sulfur source for biotin is the 2Fe-2S cluster, is BioB a "suicide enzyme",
capable of performing only a single turnover because of the destruction of the 2Fe-2S
cluster? Only two biotin-requiring proteins are present in E. coli: acetyl-CoA carboxylase
and the bifunctional biotin operon repressor/holocarboxylase synthetase (BirA)5° .
Therefore, a single turnover of multiple BioB proteins may provide enough biotin for the
- 103 -
life-cycle of E. coli. Alternatively, the 2Fe-2S cluster may be rebuilt after each turnover
by cysteine desulfurases and iron-sulfur cluster assembly proteins, conferring true
catalytic capability to BioB. Although the 2Fe-2S cluster is buried, the movement of R95
and/or Y149 could open access to the 2Fe-2S cluster from the bottom of the barrel. Thus,
the cluster may be destroyed from the top of the TIM barrel and rebuilt from the bottom.
It would not be unexpected to find that BioB requires proteins and other components to
restore its activity after turnover. The requirement of chaperones and reactivases to
repair inactivated AdoCbl-dependent radical enzymes is well documented51 .
-104-
3.5. Tables
Table 3.1. Maximal BioB activity
Cofactor content
2Fe-2S 4Fe-4S PLP
1
1
0
1
1
1
ND
Source
Mol biotin produced/
mol BioB monomer
Time (h)
Reference
0
0
1
E. coli
E. coli
E. coli
1
1
1
4
24
4
16
22
20
ND
Arabidopsis
7
6
52
'Not determined
- 105-
-
-
Table 3.2. BioB Data collection and refinement statistics
Wavelength
(A)
Resolution
range (A)
Unique
reflections
Redundancy
CompleteRdnac?o33//_<
..
I/sigma(I)
1.30000
1.73827
1.74150
100-3.7
26,018
3.4
100-4.1
100-4.5
19,002
3.1
1.1000012
100-3.4
14,030
17,465
2.8
5.4
99.5 (100.0)
98.3 (99.7)
97.1 (95.7)
98.1 (87.9)
11.8 (2.3)
14.6 (1.9)
14.0 (2.6)
15.79 (3.78)
Non-hydrogen atoms in ASU
Resolution range (A)
Number of reflections (working set / test set)
Rsym 1
6.7
6.4
8.2
6.6
(37.5)
(29.7)
(36.7)
(25.9)
4,984
44.5 - 3.4
16,222 / 1242
25.6
30.0
Rwork (%)2
Rf ee (%)
RMS deviations of protein from ideal geometry
Bonds (A)
Angles (°)
:!
Overall B (A ) (all atoms)
Group B (A2 )
AdoMet
Methionine
Ribose
Adenine
Dethiobiotin
4Fe-4S
2Fe-2S
0.009
2.3
Molecule A
87.8
Molecule B
93.3
130.5
146.3
84.6
88.6
61.8
71.0
43.0
84.6
68.3
47.4
50.8
35.7
Values in parentheses are for the highest-resolution shell. Rsy = [k a II,(hkl) <I(hkl)>l]L/l
I(hkl) for hkl independent reflections and i observations of a given
reflection, with no a cutoff. <I(hkl)> is the mean intensity of the miller index (hkl).
2 For this wavelength, Friedel pairs were merged during data processing. 3 Rwork= 'Dkl
IlFo(hkl)- IFc(hkl)ll/ kl IFo(hkl)l.No a cutoff was used in the refinement. Rfree=
Rworkfor a test set of reflections not included in refinement.
- 106-
.
Table 3.3. Metal-nitrogen (guanidine) distances from crystallographic structures of
metal-guanidine complexes.
Metal (M)
N-M distance (A)
Reference
Pt"
pt"
Pt
Co"'I
Zn"
2.06
2.07
2.018
1.94
1.95
39
39
40
39
38
-2.3-2.4
This work
l
Fe'
Table 3.4. Fe-S cluster-ligand distances'.
Atom
Contacting residue
Contacting atom
distance (A)
Cys 188
SG
2.3
Arg 260
NH1
-2.3-2.4
Fe2
Cys 97
Cys 128
SG
SG
2.2
2.2
Fel
Fe2
Cys 57
AdoMet
AdoMet
SG
0
N
2.2
-3.2
-2.4
Fe3
Fe4
Cys 60
Cys 53
SG
SG
2.2
2.2
2Fe-2S
Fel
4Fe-4S
'At our current resolution limit of 3.4 A, distances are not precisely known.
-107-
Table 3.5. Electrostatic interactions and potential hydrogen bonds between BioB and
AdoMet 1 .
AdoMet Atom
Contacting residue
Contacting atom
distance (A)
Methionine
OXT
Arg 173
NH2
2.8
Arg 173
NH2
3.2
N
N
Ribose
03*
O
Ala 100
Trp 102
0
0
3.2
3.5
Asn 153
ND2
3.5
03*
Asp 155
OD1
3.5
02*
Asn 153
ND2
3.5
02*
Asp 155
OD1
2.8
N1
Val 225
N
2.9
N6
Val 225
0
3.2
Tyr 59
0
3.3
Adenine
'At our current resolution limit of 3.4 A, distances are not precisely known.
Table 3.6. Potential hydrogen bonds between BioB and DTB
.
DTB Atom
Contacting residue
Contacting atom
distance (A)
Carboxylate
012
Thr 292
N
3.1
012
Thr 292
OG1
2.8
012
Ureido ring
0
0
N1
N2
Thr 293
N
3.0
Asn
Asn
Asn
Asn
ND2
ND2
OD1
OD1
2.9
3.7
3.7
2.8
1At our
222
153
151
222
current resolution limit of 3.4 A, distances are not precisely known.
- 108-
3.6. Figures
HN
H
NH
HN
9
6
4
BioB,DTT
Fe3+, S2 Flavodoxin
FNR, NADPH
2
Dethiobiotin (DTB)
NH
2
+NH3
NH
H
C02
Biotin
00-
(Met)
OH
OH
S-Adenosyl-L-methionine
(AdoMetor SAM)
5'-Deoxyadenosine(AdoH)
Figure 3.1. Overall reaction of BioB. BioB reconstituted anaerobically with Fe3+ and S2in the presence of DTT requires flavodoxin, flavodoxin/NADPH oxidoreductase (FNR),
and NADPH to convert DTB to biotin. Numbering schemes for DTB and AdoH are
shown. The number of equivalents of AdoMet required for the reaction is disputed (see
text).
- 109-
(a)
o
r
HNZ NH
%sfr.H
AdoMet
Met+ AdoH
[4Fe-4S]'+
[4Fe-4S]
0
NH
HN
H3C
IFe' yfe,
HHN
2C,,
H1R
NH
Met + AdoH
[4Fe-4S]2+
0
HN)
NH
HN
AdoMet
[4Fe-4S]'+
NH
HF-.sH
H
HSr
s,1- S, R
0"
Fe'
zfe I
(b)
(b)~
AdoMet
HN NH [4Fe-4S]
H
H
R
HC
Met+ AdoH
2
[4Fe-4S]
+
HN
H
e-
NH
H
H
R
S
HCH
/SH
Cys-S
NH
HN
0SH
CYS-S
Cys-S
Cys--SH
R
Cys-SH
Figure 3.2. Mechanisms proposed for the BioB reaction. (a) Mechanism put forth by
Jarrett and coworkers'6 , which requires two equivalents of AdoMet and a 2Fe-2S cluster
that is intrinsic to BioB. (b) Mechanism proposed by Fontecave and coworkers, which
requires one equivalent of AdoMet, a Cys persulfide generated from one equivalent of
Cys by a PLP-dependent Cys desulfurase activity of BioB, the presence of a thiyl radical,
and an external electron acceptor2 0 .
-110-
II
I
1
Figure 3.3. Stereo figure showing the electron density of the Fe-S clusters of BioB. The
map shown is the experimental map, after solvent flattening and two-fold NCS averaging,
at 3.7 A resolution. The map is contoured around the clusters at 3.0 a. All of the cluster
ligands (see text) are shown.
yellow, S; brown, Fe.
The atomic coloring scheme is: grey, C; red, 0; blue, N;
Figure 3.4. Solvent flattened, two-fold NCS-averaged experimental electron density of
Trp 7, contoured at 1 a. This electron density map (3.7 A resolution) was used to build
the initial model. Figure prepared with 026.
-111-
___
180
135
90
45
-45
-90
-135
Phi (degrees)
Figure 3.5. Ramachandran plot of the BioB monomer. The plot is colored as follows:
red, most favored regions; yellow, additional allowed regions; light yellow, generously
allowed regions; white, disallowed regions. Glycine residues are shown as triangles.
Figure made using PROCHECK2 9.
- 112-
(ql
(a)
_
Ih~~~~~~~~~~~~~~~~~~~~~~~~~~.
D15 5
N1 51 N1 5 3
282
R26 0
1-188
R1 7 3
1192
S2 1
VF225 *
N22 2
T2 9 2 *T 2 9 3*
-113-
Figure 3.6. Overall structure of BioB. (a) Structure of the BioB dimer. BioB exists as
a homodimer in solution'2 . We find two possible dimeric relationships between BioB
monomers in the crystal. The dimer shown here buries 17.6% of the monomer surface
area (13,249 A2 , see Fig. 3.8) and is likely to be physiologically
relevant. An
anomalous Fourier electron density map, calculated with data collected at the Fe
absorption peak wavelength (1.73827 A) and phases from the polypeptide portion of
the model, is contoured at 3 a in green mesh. These electron density peaks represent
the positions of the four Fe-S clusters in this dimeric structure. There are no other
features of similar size in the electron density map. The Fe-S cluster atoms are shown
as large spheres, with brown Fe atoms and yellow S atoms. We observe one AdoMet
(red) and one DTB (black) per subunit. All cartoon and ball-and-stick figures were
prepared with PyMOL53 , unless otherwise noted. (b) Topology diagram of the BioB
TIM barrel showing the location of important residues with respect to the -strands
(shown as arrows, numbered 1-8). The numbers above and to the left of each P-strand
correspond to the N-terminal residue of that strand; those above and to the right of the
beta strands correspond to the C-terminal residue. Ligands to the 4Fe-4S cluster are in
black, ligands to the 2Fe-2S cluster are in red, and residues that contact the 2Fe-2S
cluster ligand R260 are in green. AdoMet contacts (blue) include A100, W102, and
R173, which hydrogen bond to the amino acid moiety; D155 and N153, which
hydrogen bond to the ribose hydroxyl groups; Y59 and 1192 which stack against the
adenine ring; and V225 which forms backbone hydrogen bonds to the adenine ring.
Residues in position to hydrogen bond to DTB (brown) include N151, N153, and
N222, which contact the DTB ureido ring, T292 and T293, which contact the
carboxylate tail. Asterisks (*) denote main-chain interactions.
-114-
(a)
JA7
(b)
Figure 3.7. The BioB monomer. (a) Structure of the BioB monomer. The color
scheme is: red, 0; blue, N; yellow, S, brown, Fe. AdoMet carbon atoms are grey; DTB
carbons are black. (b) A view of the active site.
- 115-
Figure 3.8. Space filling model of BioB. Stereo figure showing the two subunits of
BioB, colored in red and blue.
-116-
Figure 3.9.
Lattice packing in BioB crystals.
Each BioB homodimer is shown in a
different color and drawn as an alpha carbon trace. Fe-S clusters are shown in brown
(Fe) and yellow (S). The BioB homodimer contacts six neighboring molecules in the
crystal. One of the BioB subunits, shown here in black, contacts a portion of
homodimers 1, 2, and 3; the other subunit, also shown in black, contacts a portion of
homodimers 1', 2', and 3'.
-
Figure 3.10. Stereo alpha carbon trace of a single BioB subunit, showing the locations of
all Lys residues in the structure. The coloring scheme is the same as that of Fig. 3.3a, but
Lys sidechains are colored in purple (C) and dark blue (No). In some cases, the Lys
sidechain was truncated due to poor electron density. K49, the only highly conserved
Lys residue in BioB, is labeled. All Lys residues are solvent-accessible and far-removed
from the active site.
-117-
(a)
o..
(b)
(c)
Figure 3.11. The BioB active site. (a) Stereo view of the 4Fe-4S cluster with AdoMet
bound. Conserved side-chain contacts between BioB and AdoMet are indicated, and
AdoMet is shown in a simulated annealing omit map contoured at 4.5a (orange). DTB is
omitted for clarity. The coloring scheme is: grey, C; red, 0; blue, N; yellow, S, brown,
Fe. (b) Stereo view of the active site, focusing on the 2Fe-2S cluster and its ligands. The
unusual R260 ligand is shown in a simulated annealing omit map contoured at 4.5a. In
addition to the 2Fe-2S cluster, R260 interacts with S43, S218, S283, and R95. Also
shown are the positions of the 4Fe-4S cluster, AdoMet, and DTB with respect to the 2Fe2S cluster. (c) Stereo view of DTB interacting with AdoMet and conserved residues
N222, N151, and N153 in the active site. Potential hydrogen bonds between N222 and
DTB are drawn as dashed lines. The stacking of the carboxylate tail of DTB and the
adenine ring of AdoMet is visible in this orientation, although contacts of the DTB
carboxylate to T292 and T293 are not (Fig. 3.15).
annealing omit map contoured at 4.0 c.
-118-
DTB is shown in a simulated
(b)
(a)
His-N
CYs-S\
Cys-S
S.,
H
His-,/,3N
I NN
FeA\;S;:
\4Fs
N
Cys--S
)__/
(C)
N
Asp
1-1
N
Fe-
S\ /
5-Cys
Fe-S\S-Cys
I\S
NH
S-|-Fe
sF!-\'S-Cys
His
Cys--S
(d)
I Fe
S-Cys
Cys-S
(e)
Cys-S
CystSO
Fe\--
S
S--
FIS
/H
Fe
H2 0
FeS
SZ--FeS
S-Cys
Cys-S
Cys
Fe-S
Cys-S
.0
0
COI
ICys
Cys
(9)
(f)
Cys-S.
S -Cys
/
Fe\Fe<
NH2
NS
Cys-S
HHN
HN Arg
Figure 3.12. 2Fe-2S and 4Fe-4S clusters with incomplete cysteinyl ligation. (a) The
Rieske protein 2Fe-2S cluster54 (b) 4Fe-4S cluster of the Fe-only hydrogenase from C.
pasteurianum. (c) 4Fe-4S cluster of the ferredoxin from P. furiosus5 5 . (d) and (e) 4Fe-4S
cluster of aconitase in the resting state and with product bound, respectively5 '5 56. (f) 2Fe2S cluster from BioB, as determined in this study. The protonation of the Arg sidechain
is inferred by analogy to small molecule metal-guanidine complexes (see Fig. 3.13), but
the true protonation state remains in question. (g) 4Fe-4S cluster from BioB, also from
this study.
- 119-
(a)
oo
N2
H'i/
(b)
H2 N
H
NH2
N
HN
HN
)
1
Ni
HN
N
NH-Zn-1 NH
CNH
H2N'
"
C_/
S
If"
I
(e)
/XN
H2N
CH 3
CH'H
NH2
Figure 3.13. Structures of aqueous metal complexes with guanidine or guanidine-like
ligands. (a) Structure of a Ni"-famotidine complex, proposed from spectroscopic
studies46 . (b) Crystallographic structure of a Zn" complex with a guanidine ligand4 7 . (c)
Crystallographic structure of a Co"' complex with a guanidine ligand4 8 . (d) and (e)
.
Crystallographic structures of Pt" complexes with guanidine ligands
BioB
LipA
53-CPEDCKYC
94-CTRRCPFC
KamA 125-CSMYCRHC
Pf1A 30-CLMRCLYC
192-IVGLGE
217-MVGLGE
262-LRGVND
171-VPGWSD
Figure 3.14. Conservation of residues across the AdoMet radical enzymes. Enzymes are
denoted by their corresponding gene names (BioB, E. coli biotin synthase; LipA, E. coli
lipoyl synthase; KamA, C. subterminale lysine 2,3-aminomutase; PflA, E. coli PFL
activase). Absolutely conserved residues, including the CxxxCxxC motif that ligates the
4Fe-4S cluster, are shown in bold red typeface. Bold black typeface denotes similar
residues. Y59 in BioB, which forms 7e-stackinginteractions with AdoMet, corresponds
to an aromatic residue in the other enzymes. 1192 forms van der Waal contacts to
AdoMet and corresponds to a hydrophobic residue in the other enzymes.
- 120-
T292
II
T293
Figure 3.15. The carboxylate tail of DTB interacting with residues T292 and T293.
These two residues are part of the loop that follows strand 8 of the TIM barrel. The
coloring scheme is: grey, C; blue, N; red, 0.
- 121 -
3.7. References
1.
2.
3.
4.
Sofia, H.J., Chen, G., Hetzler, B.G., Reyes-Spindola, J.F. & Miller, N.E. Radical
SAM, a novel protein superfamily linking unresolved steps in familiar
biosynthetic pathways with radical mechanisms: functional characterization using
new analysis and information visualization methods. Nucleic Acids Res. 29, 10971106 (2001).
Krebs, C., Broderick, W.E., Henshaw, T.F., Broderick, J.B. & Huynh, B.H.
Coordination of adenosylmethionine to a unique iron site of the [4Fe-4S] of
pyruvate formate-lyase activating enzyme: A mossbauer spectroscopic study.
Journal of the American Chemical Society 124, 912-913 (2001).
Walsby, C.J., Ortillo, D., Broderick, W.E., Broderick, J.B. & Hoffman, B.M. An
anchoring role for FeS clusters: chelation of the amino acid moiety of Sadenosylmethionine to the unique iron site of the [4Fe-4S] cluster of pyruvate
formate-lyase activating enzyme. Journal of the American Chemical Society 124,
11270-11271 (2002).
Cosper, M.M. et al. The [4Fe-4S]2+ cluster in reconstituted biotin synthase binds
S-adenosylmethionine. Journal of the American Chemical Society 124, 1400614007 (2002).
5.
Chen, D., Walsby, C.J., Hoffman, B.M. & Frey, P.A. Coordination and
mechanism of reversible cleavage of S-adenosylmethionine by the [4Fe-4S]
center in lysine 2,3-aminomutase. Journal of the American Chemical Society 125,
6.
Berkovtich, F., Nicolet, Y., Wan, J.T., Jarrett, J.T. & Drennan, C.L. Crystal
structure of biotin synthase, an S-adenosylmethionine-dependent radical enzyme.
Science 303, 76-79 (2004).
Escalletes, F., Florentin, D., Tse Sum Bui, B., Lesage, D. & Marquet, A. Biotin
Synthase Mechanism: Evidence for Hydrogen Transfer from the Substrate into
Deoxyadenosine. Journal of the American Chemical Society 121, 3571-3578
11788-11789 (2003).
7.
(1999).
8.
Benda, R. et al. Iron-Sulfur Clusters of Biotin Synthase In Vivo: A Mossbauer
9.
Study. Biochemistry 41, 15000-15006 (2002).
Duin, E.C. et al. [2Fe-2S] to [4Fe-4S] cluster conversion in Escherichia coli
biotin synthase. Biochemistry 36, 11811-11820 (1997).
10.
11.
Ollagnier-de Choudens, S. et al. Iron-sulfur center of biotin synthase and lipoate
synthase. Biochemistry 39, 4165-4173 (2000).
Ollagnier-de Choudens, S., Mulliez, E., Hewiston, K.S. & Fontecave, M. Biotin
synthase is a pyridoxal phosphate-dependent cysteine desulfurase. Biochemistry
41, 9145-9152 (2002).
12.
13.
Sanyal, I., Cohen, G. & Flint, D.H. Biotin synthase: purification, characterization
as a [2Fe-2S]cluster protein, and in vitro activity of the Escherichia coli bioB
gene product. Biochemistry 33(1994).
Tse Sum Bui, B., Florentin, D., Marquet, A., Benda, R. & Trautwein, A.X.
Mossbauer studies of Escherichia coli biotin synthase: evidence for reversible
- 122 -
14.
15.
interconversion between [2Fe-2S]2+ and [4Fe-4S]2 + clusters. Federation of
European Biochemical Societies Letters 459, 411-414 (1999).
Ugulava, N.B., Gibney, B.R. & Jarrett, J.T. Biotin synthase contains two distinct
iron-sulfur cluster binding sites: chemical and spectroelectrochemical analysis of
iron-sulfur interconversions. Biochemistry 40, 8343-8351 (2001).
Ugulava, N.B., Surerus, K.K. & Jarrett, J.T. Evidence from Mossbauer
spectroscopy for distinct [2Fe-2S](2+) and [4Fe-4S](2+) cluster binding sites in
biotin synthase from Escherichia coli. J. Am. Chem. Soc. 124, 9050-1 (2002).
16.
17.
Ugulava, N.B., Sacanell, C.J. & Jarrett, J.T. Spectroscopic Changes during a
Single Turnover of Biotin Synthase: Destruction of a [2Fe-2S] Sluster
Accompanies Sulfur Insertion. Biochemistry 40, 8352-8358 (2001).
Tse Sum Bui, B. et al. Biotin synthase mechanism: on the origin of sulphur.
Federationof EuropeanBiochemicalSocietiesLetters440, 226-230(1998).
18.
19.
20.
21.
Tse Sum Bui, B. et al. Fate of the (2Fe-2S)2+ Cluster of Escherichia coli Biotin
Synthase during Reaction: A M6ssbauer Characterization. Biochemistry 42, 87918798 (2003).
Florentin, D., Bui, B.T., Marquet, A., Ohshiro, T. & Izumi, Y. On the mechanism
of biotin synthase of Bacillus sphaericus. Comptes Redus de L Academie des
Sciences Serie III - Sciences de la Vie 317, 485-488 (1994).
Ollagnier-de Choudens, S., Mulliez, E. & Fontecave, M. The PLP-dependent
biotin synthase from Escherichia coli: mechanistic studies. Federation of
European Biochemical Societies Letters 532, 465-468 (2002).
Cosper, M.M. et al. Characterization of the cofactor composition of Escherichia
coli biotin synthase. Biochemistry 43, 2007-2021 (2004).
22.
Jameson, G.N., Cosper, M.M., Hernandez, H.L., Johnson, M.K. & Huynh, B.H.
Role of the [2Fe-2S] cluster in recombinant Escherichia coli biotin synthase.
Biochemistry 43, 2022-2031 (2004).
23.
Terwilliger, T.C. & Berendzen, J. Automated structure solution for MIR and
MAD. Acta Crystallographica D55, 849-861 (1999).
24.
Kabsch, W. A solution for the best rotation to relate two sets of vectors. Acta
Crystallographica A32, 922-923 (1976).
25.
26.
27.
Cowtan, K.D. & Zhang, K.Y.J. Density modification for macromolecular phase
improvement. Progress in Biophysics and Molecular Biology 72, 245-270 (1999).
Jones, T.A., Zou, J.-Y., Cowan, S.W. & Kjeldgaard, M. Improved methods for
building protein models in electron density maps and the location of errors in
these models. Acta Crystallographica A47, 110-119 (1991).
Brunger, A.T. et al. Crystallography & NMR system: A new software suite for
macromolecular structure determination. Acta Crystallographica D55, 905-921
(1998).
28.
29.
30.
Collaborative Computational Project Number 4. The CCP4 suite: programs for
protein crystallography. Acta Crystallographica D50, 760-763 (1994).
Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton, J.M. PROCHECK:
a program to check the stereochemical quality of protein structures. Journal of
Applied Crystallography 26, 283-291 (1993).
Layer, G., Moser, J., Heinz, D.W., Jahn, D. & Schubert, W.D. Crystal structure of
coproporphyrinogen III oxidase reveals cofactor geometry of Radical SAM
- 123-
31.
32.
33.
enzymes. European Molecular Biology Organization Journal 22, 6214-6224
(2003).
Murzin, A.G., Brenner, S.E., Hubbard, T. & Chothia, C. SCOP: a structural
classification of proteins database for the investigation of sequences and
structures. Journal of Molecular Biology 247, 536-540 (1995).
Cosper, M.M., Jameson, G.N.L., Eidsness, M.K., Huynh, B.H. & Johnson, M.K.
Recombinant Escherichia coli biotin synthase is a [2Fe-2S]2+ protein in whole
cells. Federation of European Biochemical Societies Letters 529, 332-336 (2002).
Hewitson, K.S., Baldwin, J.E., Shaw, N.M. & Roach, P.L. Mutagenesis of the
proposed iron-sulfur cluster binding ligands in Escherichia coli biotin synthase.
Federationof EuropeanBiochemicalSocietiesLetters 466, 372-376(2000).
34.
35.
36.
Ollagnier-de Choudens, S., Sanakis, Y., Hewiston, K.S., Roach, P. & Munck, E.
Reductive cleavage of S-Adenosylmethionine by biotin synthase from
Escherichia coli. Journal of Biological Chemistry 277, 13449-13454 (2002).
Ugulava, N.B., Gibney, B.R. & Jarrett, J.T. Iron-Sulfur cluster Interconversion in
Biotin Synthase: Dissociation and Reassociation of Iron during Conversion of
[2Fe-2S] to [4Fe-4S] Clusters. Biochemistry 39, 5206-5214 (2000).
Tse Sum Bui, B. et al. Fate of the (2Fe-2S)2+ cluster of Escherichia coli biotin
synthase during reaction: a Mossbauer characterization. Biochemistry 42, 87918798 (2003).
37.
38.
Hewitson, K.S. et al. The iron-sulfur center of biotin synthase: site-directed
mutants. Journal of Biological Inorganic Chemistry 7, 83-93 (2002).
Cosper, M.M. et al. The [4Fe-4S]2+ cluster in reconstituted biotin synthase binds
S-adenosyl-L-methionine.
39.
J. Am. Chem. Soc. 124, 14006-14007 (2002).
Liu, A. & Graslund, A. Electron paramagnetic resonance evidence for a novel
interconversion of [3Fe-4S](+) and [4Fe-4S](+) clusters with endogenous iron and
sulfide in anaerobic ribonucleotide reductase activase in vitro. J. Biol. Chem. 275,
12367-73 (2000).
40.
41.
Padovani, D., Thomas, F., Trautwein, A.X., Mulliez, E. & Fontecave, M.
Activation of class III ribonucleotide reductase from E. coli. The electron transfer
from the iron-sulfur center to S-adenosylmethionine. Biochemistry 40, 6713-9
(2001).
Cosper, M.M. et al. Structural studies of the interaction of S-adenosylmethionine
with the [4Fe-4S] clusters in biotin synthase and pyruvate formate-lyase
activating enzyme. Protein Science 12, 1573-1577 (2003).
42.
43.
44.
Schubert, H.L., Blumenthal, R.M. & Cheng, X. Many paths to methyltransfer: a
chronicle of convergence. Trends in Biochemical Sciences 28, 329-335 (2003).
Hoffman, J.L. Chromatographic analysis of the chiral and covalent instability of
S-adenosyl-L-methionine. Biochemistry 25, 4444-4449 (1986).
de la Haba, G., Jamieson, G.A., Harvey Mudd, S. & Richards, H.H. Sadenosylmethionine: the relation of configuration at the sulfonium center to
enzymatic reactivity. Journal of the American Chemical Society 81, 3975-3980
(1959).
45.
Ugulava, N.B., Frederick, K.K. & Jarrett, J.T. Control of AdenosylmethionineDependent Radical Generation in Biotin Synthase: A Kinetic and Thermodynamic
- 124-
46.
Analysis of Substrate Binding to Active and Inactive Forms of BioB.
Biochemistry 42, 2708-2719 (2003).
Baranska, M., Gumienna-Kontecka, E., Kozlowski, H. & Proniewicz, L.M. A
study on the nickel(II)-famotidine complexes. Journal of Inorganic Biochemistry
92, 112-120 (2002).
47.
Aoki, S., Iwaida, K., Hanamoto, N., Shiro, M. & Kimura, E. Guanidine is a Zn 2+ -
48.
binding ligand at neutral pH in aqueous solution. Journal of the American
Chemical Society 124, 5256-5257 (2002).
Fairlie, D.P. et al. Models for Arginine-Metal Binding. Synthesis of Guanidine
and Urea Ligands through Amination and Hydration of a Cyanamide Ligand
Bound to Platinum (II), Osmium (III), and Cobalt (III). Inorganic Chemistry 36,
1020-1028 (1997).
49.
50.
51.
52.
Ratilla, E.M.A., Scott, B.K., Moxness, M.S. & Kostic, N.M. Terminal and new
bridging coordination of methylguanidine, arginine, and canavanine to
platinum(II). The first crystallographic study of bonding between a transition
metal and a guanidine ligand. Inorganic Chemistry 29, 918-926 (1990).
Wilson, K.P., Shewchuk, L.M., Brennan, R.G., Otsuka, A.J. & Matthews, B.W.
Escherichia coli biotin holoenzyme synthetase/bio repressor crystal structure
delineates the biotin- and DNA-binding domains. Proceedings of the National
Academy of Sciences 89, 9257-9261 (1992).
Toraya, T. Radical catalysis in coenzyme B12 -dependent isomerization
(eliminating) reactions. Chemical Reviews 103, 2095-2127 (2003).
Picciocchi, A., Douce, R. & Alban, C. Biochemical characterization of the
Arabidopsis biotin synthase reaction. The importance of mitochondria in biotin
synthesis. Plant Physiol. 127, 1224-1233 (2001).
53.
54.
DeLano, W.L. The PyMOL Molecular Graphics System. (2002).
Colbert, C.L., Coutre, M.M., Eltis, L.D. & Bolin, J.T. A cluster exposed: structure
of the Rieske ferredoxin from biphenyl dioxygenase and the redox properties of
Rieske Fe-S proteins. Structure 8, 1267-1278 (2000).
55.
Zhou, Z.H. & Adams, M.W.W. Site-directed mutations of the 4Fe-ferredoxin
from the hyperthermophilic archaeon Pyrococcusfuriosus: role of the clustercoordinating aspartate in physiological electron transfer reactions. Biochemistry
36, 10892-10900 (1997).
56.
Zheng, L., Kennedy, M.C., Beinert, H. & Zalkin, H. Mutational analysis of active
site residues in pig heart aconitase. Journal of Biological Chemistry 267, 78957903 (1992).
- 125-
Chapter 4
Crystal structure of lysine 5,6-aminomutase
- 126-
Abbreviations:
2.3-LAM
Lysine 2,3-aminomutase
5,6-LAM
Lysine 5,6-aminomutase
Ado.
AdoCbl
AdoH
AdoMet
Cbl
CNCbl
CoA
Cob(II)
DMB
GM
MCM
MS
PEG
SAD
SeMet
OAM
PLP
TIM
5'-deoxyadenosyl radical
Adenosylcobalamin, Coenzyme B 12
5'-deoxyadenosine
S-adenosyl-L-methionine
Cobalamin
Cyanocobalamin, Vitamin B12
Coenzyme A
Cob(II)alamin
Dimethylbenzimidazole
Glutamate mutase
Methylmalonyl coenzyme A mutase
Methylcobalamin-dependent methionine synthase
Polyethylene glycol
Single-wavelength anomalous dispersion
Selenomethionine
Omithine aminomutase
Pyridoxal 5'-phosphate
Triosephosphate isomerase
- 127 -
Introduction
Lysine 5,6-aminomutase (5,6-LAM) is an enzyme in the bacterial lysine fermentation
pathway.
The enzyme catalyzes the interconversion of DL-lysine or of f3-L-lysine to
2,5-diaminohexanoate or to 3,5-diaminohexanoate, respectively (Fig. 4.1). The crystal
structure of 5,6-LAM, a major contribution to this thesis, was determined in collaboration
with the laboratory of P. A. Frey at the University of Wisconsin at Madison. We present
a crystal structure of a substrate-free holoenzyme form of 5,6-LAM from Clostridium
sticklandii, which is the first structure of an enzyme that utilizes both adenosylcobalamin
(AdoCbl) and pyridoxal 5'-phosphate (PLP) cofactors. The structure reveals that AdoCbl
is bound by a Rossmann-like domain, and that PLP, which is tethered to the B 12 -binding
domain via its imine linkage, is bound in the putative active site, at the top of a
triosephosphate isomerase (TIM) barrel domain. Thus, 5,6-LAM joins a group of three
other AdoCbl-dependent radical enzymes (glutamate mutase [GM], Methylmalonyl-CoA
mutase [MCM], and diol dehydratase, see chapter 1)24 and one adenosylmethionine
(AdoMet) -dependent radical enzyme (biotin synthase, see chapter 3)5 in utilizing the
TIM barrel fold to sequester substrates that form free-radical intermediates. Our structure
is unique among all PLP-dependent enzymes, in that the imine linkage is provided by a
Rossmann-like domain.
Strikingly, in the pre-catalytic, substrate-free conformation
captured in our structure, the PLP and AdoCbl cofactors are separated by a distance of
-25A, suggesting that a gross conformational change occurs upon substrate binding. The
work presented in this chapter is in press at the Proceedings of the National Academy of
Sciences 6 .
- 128 -
AdoCbl-dependent isomerases are often present in catabolic pathways and can serve to
rearrange the substrate's carbon skeleton and/or functional groups for further degradation.
One such pathway that operates in several bacterial species is the fermentation of lysine
to yield acetate. Interestingly, the lysine fermentation pathway contains two analogous
enzymes: 5,6-LAM, which is AdoCbl-dependent7 8 , and lysine 2,3-aminomutase (2,3LAM), which is an AdoMet-dependent radical enzyme9 ' l . Both enzymes require PLP9,'12
in addition to AdoCbl or AdoMet, and both catalyze a 1,2 amino group shift with
concomitant H atom migration (Fig. 4.1). In 5,6-LAM, AdoCbl is the source of the
transient 5'-deoxyadenosyl radical (Ado.), whereas in 2,3-LAM, Ado* is produced from
AdoMet coordinated to the unique iron of a 4Fe-4S cluster, and an exogenous electron.
The exogenous electron is transferred first into the [4Fe-4S]2+ cluster and then into
AdoMet, cleaving AdoMet to give methionine and the transient Ado..
Because of the
similarity of their reactions, their requirement for PLP, and their common substrates, 5,6LAM may be thought of as the AdoCbl-dependent counterpart of the AdoMet-dependent
2,3-LAM. Studies on both 2,3-LAM and 5,6-LAM have allowed Frey and coworkers to
propose a general mechanism for these reactions (Fig. 4.2a). For both 5,6-LAM and 2,3LAM, radical propagation from Ado. to the substrate-PLP covalent complex (known as
the external aldimine) initiates the isomerization, and both reaction mechanisms are likely
to involve analogous intermediates.
The role of PLP in the aminomutase reaction is not completely clear. The 5,6-LAM
mechanism proposed by Frey and coworkers is consistent with computational studies
conducted by Random's laboratory, which suggest that the role of PLP in the AdoCbl-
- 129-
dependent aminomutases is to provide a conjugated n system to stabilize the rearranging
radical species1 3 14. This hypothesis is consistent with model chemistry, wherein an imine
with an aromatic substituent undergoes a 1,2-imino shift under radical generating
conditions 15 (Fig. 4.3a). Significantly, the rearrangement of the aziridylcarbinyl radical,
which is the simplest model for the 5,6-LAM
reaction, takes place at cryogenic
temperatures, despite the absence of an aromatic substituent' 6 (Fig. 4.3b), indicating that
PLP is not necessary for the rearrangement to occur.
All AdoMet- or AdoCbl-dependent radical enzymes rely on Ado. for catalysis, yet the
formation of this highly oxidative intermediate must be controlled in order to prevent
aberrant reactions.
C-Co bond homolysis and the transient formation of Ado
is
triggered by substrate binding in the AdoCbl dependent enzymes MCM' 7 , GM' 8 '19, and
diol dehydratase2 0 , while effector binding triggers the C-Co bond homolysis in the
AdoCbl-dependent ribonucleotide reductase21 (see chapter 1).
How substrates or
effectors afford a 1012 -fold rate acceleration of Co-C bond cleavage in AdoCbl-dependent
enzymes is a question that has intrigued scientists for decades. The crystal structure of
5,6-LAM suggests a novel role for PLP in locking the enzyme into its resting-state
conformation, keeping AdoCbl out of the active site in the absence of substrate and
preventing radical-based protein damage. This will be discussed in detail below.
- 130-
4.2. Materials and methods
4.2.1.Proteinpreparation
Selenomethionine (SeMet)-incorporated 5,6-LAM from Clostridium sticklandii was
purified and reconstituted with PLP in the Frey laboratory.
A plasmid KamDE
containing the genes encoding both [3and a subunits of 5,6-LAM was used to transform
E. coli B834 (DE3) cells. Ten mL of an overnight culture was used to inoculate 1 L of
SeMet media, prepared according Budisa and coworkers22 with the omission of Lcysteine and addition of 5 M pyridoxal hydrochloride. The culture was grown for 16
hours at 37
C. The yield was 25 g wet cells from 10 L minimal media.
The SeMet
recombinant protein was purified by the procedure of Chang and Frey23 without the gel
filtration chromatography step.
4.2.2. Storage, handling, and crystallization of protein samples, and crystal
cryoprotection
Protein was stored at -80 C and handled on ice during crystallization. All solutions
containing AdoCbl (including protein solutions) were formulated and handled in a dark
room under dim red light.
Crystals were grown using hanging drop vapor diffusion techniques at room temperature
in a dark room, under dim red light. Crystals were manipulated in the dark room until
after cryo-cooling. Five crystal forms were obtained, as detailed in Fig. 4.4. Of these
crystals, only one crystal form (form 1, Fig. 4.4a) produces data adequate to solve the
structure. Protein and precipitant solutions (Fig. 4.4) were mixed in a 1:1 ratio, with drop
- 131 -
sizes of 2 gL total, and equilibrated over 0.5 mL of precipitant solution. Form 1 crystals
(those used to solve the structure) were cryoprotected, within 48 hours after their initial
appearance, by quickly soaking in precipitant solution with 20% glycerol added and
plunging into liquid nitrogen.
4.2.3. Crystalproperties,phasing, model building,and refinement
Crystals belong to space group P3121, with one ad heterodimer per asymmetric unit, for
a total of 6 aouheterodimers in the unit cell (Z=6), and a = b = 99.7 A, c = 168.8 A. The
unit cell volume V = 1,453,093 A3 . Data were collected at the Argonne National
Laboratory beamline NE-CAT 8-BM, equipped with a charge-coupled device detector
(Area Detector Systems Corp.). Data were processed and scaled with DENZO and
SCALEPACK 2 4 .
A single set of data, collected at the Se absorption peak wavelength (0.97918 A), was
used to solve the structure by anomalous dispersion methods2 '52 6 (Table 4.1). Twentynine Se atoms and the Co atom of AdoCbl were located and refined with SOLVE2 7 , with
a mean figure of merit of 0.40 to 3.0 A resolution. The experimental electron density
map was subjected to solvent-flattening with RESOLVE2 7 , resulting in a good-quality
map that allowed us to begin model building (Fig. 4.5). The space group, P3 121, was
distinguished from P32 21 by inspection of a helices, which are right-handed.
Iterative rounds of model building in XFIT2 8 and refinement in CNS2 9 resulted in the
final model at 2.8
A resolution.
The refined structure contains all 516 residues of the a
- 132 -
subunit (chain A), residues 24-84 and 102-261 of the
P subunit (chain B), one AdoCbl
molecule, and one PLP molecule (as the imine adduct to K144p). A simulated annealing
composite omit map was used to validate the final structure. There is no electron density
for the histidine tag, residues 1-231, 85-101p, or 262P. The final model has all residues
residing in the allowed regions of the Ramachandran plot (87.7% in the most favored
regions, 11% in additionally allowed regions, and 1.3% in generously allowed regions),
as calculated by PROCHECK 3 0 (Fig. 4.6).
- 133-
4.3. Results
4.3.1. Overallstructure
5,6-LAM is an a2 P2 tetramer3l that can be thought of as a dimer of a3 units (Fig. 4.7). In
the crystal, the asymmetric unit contains one a3 dimer, and crystallographic symmetry
produces a likely physiological tetramer that buries 5736 A2 (19%) of the ap heterodimer
surface area (Fig. 4.9). The large a subunit (538 residues) is composed of the PLPbinding TIM barrel domain and several additional a helices and P strands at the N and C
termini (Fig. 4.8).
These helices and strands form an intertwined "accessory clamp"
structure that wraps around the sides of the TIM barrel, and extends up toward the Ado
ligand of the Cbl cofactor. This accessory clamp provides most of the interactions
between the protein and the Ado ligand of the Cbl (discussed below), suggesting that its
role is mainly in stabilizing AdoCbl in the pre-catalytic resting state. The small [5subunit
(262 residues) comprises two domains: the N-terminal dimerization domain, which has
the same fold as copper binding domain of the Alzheimer's disease amyloid precursor
protein32 (PDB code 1OWT), and the AdoCbl-binding Rossmann-like domain, which
also provides the imine bond to PLP. The Rossmann-like domain interacts with the Cterminal end of the TIM barrel, placing PLP into the top of the barrel, while projecting
AdoCbl to the edge of the barrel, far from the PLP binding site. The dimerization domain
of p forms a continuous n-sheet with the dimerization domain of the second a[ unit and
buries an extensive surface of hydrophobic residues.
The dimerization domain
also makes many hydrogen-bonding and hydrophobic contacts with the TIM barrel
domain of the a subunit. No contacts are observed between the dimerization domain and
the Rossmann domain, though these domains are linked by a disordered loop (85Bf-101f)
- 134-
that we could not model.
The average B-factor of the dimerization domain is
approximately twice those of either the Rossmann-like domain or the TIM barrel and
accessory clamp, suggesting a degree of mobility for the dimerization domain (Table
4.1). The a 2 P2 tetramer is arranged such that there is no direct interaction between the
active sites of either ap pair.
The presence of a TIM barrel and a Rossmann-like domain in 5,6-LAM is consistent with
the domain usage in other Cbl-dependent base-off enzymes, such as GM, MCM, and
methionine synthase (MS), but the orientation of the Rossmann domain relative to the
TIM barrel is markedly different, and is likely to be mechanistically relevant (discussed
below).
It is also interesting that both subunits play a role in binding both cofactors;
AdoCbl is bound at the C-terminal end of the Rossmann domain of
C
and by the
accessory clamp at the edge of the a subunit, whereas PLP is bound at the C-terminal end
of the TIM domain of a and by a lysine residue at the edge of the P subunit.
4.3.2. PLP-proteininteractions
PLP-dependent enzymes of known structure can be sorted into families based on their
respective folds. Enzymes of fold type I, II, III, IV, and V are represented by aspartate
aminotransferase,
tryptophan
synthase,
alanine
racemase,
D-amino
aminotransferase, and glycogen phosphorylase, respectively3 3 (Fig. 4.10).
acid
5,6-LAM
cannot be placed into any of these five PLP enzyme families, although it does share
features with fold types II, III, and IV. As with fold type III, PLP is bound at the top of
the TIM barrel pore, but the imine linkage is formed to the Rossmann domain rather than
135 -
to a lysine of the TIM barrel, and a least-squares superposition of 5,6-LAM and alanine
racemase reveals that PLP does not occupy analogous positions at the C-terminal ends of
the TIM barrels (Fig. 4.11). In addition, the second domain of fold type III is composed
mainly of P strands rather than having a Rossmann-like fold. The similarity to fold type
II proteins is more limited and focuses on the use of a serine residue (S238a in 5,6-
LAM), rather than the more typical aspartic acid, to hydrogen bond with the pyridine
nitrogen. The presence of a Ser residue in this position suggests that the pyridine
nitrogen is not protonated in 5,6-LAM. In agreement with fold types III and IV, the si
face of PLP is solvent exposed3 3 .
The experimental electron density map indicates that K144j3 forms a linkage to PLP (Fig.
4.5). Biochemical, mutagenesis and mass-spectrometry studies on 5,6-LAM from P.
gingivalis3 4 , combined with sequence alignment against the C. sticklandii enzyme (-67%
identity), support the assignment of K144P as the Lys residue that forms an imine bond
with PLP.
K1441 resides at the N-terminus
of a short glycine-rich
loop (1441-
KGYAGHYG-15113) that is highly conserved across all 5,6-LAM sequences.
This short
loop interrupts the second helix of the Rossmann domain (Figs. 4.8, 4.12b). The helix is
reformed for one last turn before the peptide resumes the remainder of the Rossmann
fold. As was previously recognized3 4 , this represents a novel PLP-binding motif. The
terminal amino group of K14413does not appear to be coplanar with the pyridine ring of
PLP, leading us to believe that the internal aldimine was photoreduced in the X-ray beam,
as has been observed in other structures (e.g. PDB codes 1GOP3 5, 1B5436, 1T313 7 ).
Significantly, all contacts to the PLP cofactor, except for the imine linkage, are made
136 -
from residues of the TIM barrel (Fig. 4.12a and Table 4.2), and are similar to PLP-protein
interactions observed in other PLP-dependent enzymes. These interactions include n7stacking, electrostatic interactions with the phosphate, and hydrogen bonding. Y263a
forms n-electron stacking interactions with the pyridine ring and also hydrogen bonds to
the phosphate moiety. The phenolic oxygen of PLP, proposed to be important for
intermediate
stabilization
in PLP-
and AdoCbl-dependent
1,2-aminomutases 1 4 , is
observed to hydrogen bond to the side-chain of N299a. The phosphate group interacts
with two Arg sidechains (R184a, R268a), the sidechain of S189a, and a number of
main-chain amides (G187a, Q188a, S189a).
Like in all PLP enzymes of known
structure, the phosphate moiety of PLP is bound near the N-terminus of an a helix (in this
case, a short helical turn, composed of 188a-QSL-190a).
4.3.3. Cbl-protein interactions
AdoCbl binds to 5,6-LAM with KM= 6.6 CiMand stabilizes the enzyme against thermal
lability2 3 . AdoCbl was added to 5,6-LAM immediately prior to crystallization, and we
observe electron density that is consistent with the cofactor (Fig. 4.13a,b). 5,6-LAM
contains a "base-off' AdoCbl binding sequence (131P5-DxHxxG...Sxl...GG-222[3)and
binds AdoCbl in the base-off conformation, with H133P replacing the intrinsic
dimethylbenzimidazole
(DMB) substituent of the cofactor as the lower axial ligand to the
cobalt23 (the Co-N distance is 2.3 A). Binding of AdoCbl by the Rossmann domain of
5,6-LAM is similar to Cbl binding in MS38 , MCM 3 , and GM2 (see chapter 1), and is
characterized by a hydrogen-bonding network designed to bind the Cbl cofactor (Fig
4.13a, b, and Table 4.3). The binding determinants for Cbl include several residues that
- 137-
hydrogen bond to the propionamide side chains of the corrin ring (T130p, A132p, T134P,
V135[, T1911, Q192P), and, as in the case of MS, MCM, and GM, a serine residue
(S1871) which hydrogen-bonds to the DMB N atom. One potential hydrogen bond
between the ribose moiety of the DMB tail and the protein is observed: the main-chain
carbonyl O of R243P is 2.8 A away from the 2'-OH.
The DMB moiety is bound in a
largely hydrophobic cavity, and is in van der Waal contact with I137/5, I140/5, V248p,
L1853, and F23913, as well as G221[ and G2221, two residues of the base-off Cbl
binding sequence motif. The phosphate moiety of the DMB tail, which is not observed to
contact the protein directly, is bound near the surface of the Rossmann-like domain.
Based on the structures of GM and MCM, we assume that the phosphate interacts with
solvent, but due to the moderate resolution of our structure, we did not model water
molecules.
All solutions and crystals containing AdoCbl were handled under red light until after
cryo-cooling. Despite our stringent efforts to prevent the cleavage of the photolabile CCo bond, the simulated annealing composite omit map suggest that this bond is mostly
cleaved (Fig. 4.13b), and we have modeled the Ado moiety as AdoH. We believe that the
cleavage of AdoCbl is nonenzymatic and is due to photoreduction of the C-Co bond in
the X-ray beam. This phenomenon has previously been reported39 . In the structure of
5,6-LAM, AdoH adopts a syn conformation about the glycosidic bond, in contrast to
GM40 and MCM41 , where the adenine ring of Ado is anti to the ribose ring. Interestingly,
AdoH is forced to adopt the syn conformation because of a steric clash between the
adenine ring and Y193a that would arise if AdoH were in the anti conformation. AdoH
- 138-
interacts mostly with residues of the accessory clamp: the 2' and 3' -OH groups of the
ribose moiety of AdoH hydrogen bond to the main-chain carbonyls of E55a and D54a,
respectively, and the exocyclic amino group of the adenine ring hydrogen bonds to D64ac
Y193a and V56a are positioned for hydrophobic interaction with the adenine moiety of
AdoH. Interestingly, the AdoH binding loop, composed of residues 51a-57a, form a
distorted P-hairpin structure that allows for two adjacent main-chain carbonyls to
hydrogen bond with the 2'- and 3'-OH groups of AdoH (Fig. 4.13b, Table 4.3).
4.3.4. The CxxCxxxC motif of 5,6-LAM
Interestingly, 5,6-LAM contains a CxxCxxxC motif and is reported to adventitiously bind
metals23 . Analysis of the 5,6-LAM structure suggests that the three-Cys sequence, which
is reminiscent of the conserved CxxxCxxC motif of the AdoMet radical enzymes, plays
no role in catalysis. The three Cys residues are not positioned for disulfide formation or
metal binding (Fig. 4.14), and there is no additional electron density to suggest the
presence of a metal in the area.
Moreover, these Cys residues are not absolutely
conserved across the sequenced 5,6-LAMs.
-139-
4.4 Discussion
4.4.1. A hypothetical conformational change to bring AdoCbl into the active site
As mentioned above, the relative orientation of the Rossmann-like domain to the TIM
barrel in 5,6-LAM is markedly different from that observed in either GM or MCM.
Whereas in 5,6-LAM the Rossmann domain is bound off-center on top of the TIM barrel
(Fig. 4.15a), resulting in the -25A separation of the AdoCbl cofactor from the active-site
PLP, in the enzyme-substrate complex of both MCM and GM, the Rossmann domain is
docked directly over the center of the TIM barrel, thereby placing the Cbl cofactor in the
active site and sequestering the active site from bulk solvent2'3 . MCM buries -15% and
-60% of the TIM-barrel and Rossmann domain surface areas, respectively, at the TIM
barrel-Rossmann interface. Likewise, GM buries -17.5% of the TIM barrel surface area
and -47% of the Rossmann domain surface area. In contrast, 5,6-LAM buries only
-6.6% of the TIM barrel domain and 17% of the Rossmann domain at the TIM barrelRossmann interface. Therefore, we describe this structure of 5,6-LAM, in which the
active site is solvent-accessible and the AdoCbl cofactor is not positioned for catalysis, as
having the resting state "edge-on" conformation, in contrast to the hypothetical catalytic
"top-on" conformation (Fig. 4.15b) that is analogous to the domain arrangement in the
substrate-bound forms of MCM and GM (Fig. 4.15d).
A key feature that is unique to PLP binding in the resting-state structure of 5,6-LAM is
the intersubunit nature of the imine linkage (Fig. 4.12). In this mode of PLP binding, the
cofactor acts as an anchor, tethering the separate polypeptide chain of the Rossmann-like
domain to the TIM barrel domain through PLP. We propose that the anchoring role of
- 140-
PLP, which forces the unusual orientation of the Rossmann domain with respect to the
TIM barrel, is important for positioning AdoCbl outside of the putative active site and for
regulating the formation of the highly oxidizing Ados in the absence of substrate. In the
resting state, a large cleft separates the dimerization domain from the Rossmann-like
domain (Fig. 4.16). This cleft, which leads to the top of the TIM barrel and the PLP
cofactor, is presumably the path that the substrate must follow in order to arrive at the
active site. Introduction of the substrate and transaldimination would then release K144P
and effectively break the PLP-mediated anchoring interactions between the Rossmannlike and TIM barrel domains. With the Rossmann-like domain freed from its constrained
position, a large-scale conformational change could occur.
We suggest that such a
conformational change must occur, and must result in the docking of the Rossmann-like
domain directly atop the center of the TIM barrel domain, filling the cleft between the
Rossmann domain and the dimerization domain (Fig. 4.15b). This catalytic "top-on"
conformation would effectively sequester the active site and position Ado near the
substrate-PLP complex, allowing for substrate radical generation and catalysis. Such a
conformational change could provide the energy that is required for the large rateacceleration of AdoCbl C-Co bond homolysis that characterizes this enzyme family. 5,6LAM is not exceptionally fast (kcat= 750 +44 min1- for D-lysine4 2 ), and would not require
a very fast conformational change for catalysis. The hypothetical conformational change
that accompanies formation of the catalytic state of 5,6-LAM could not be rate-limiting,
since a primary deuterium kinetic isotope effect is observed in the reaction of 5,6-LAM
with deuterium labeled substrates.
- 141 -
There is precedent in the Cbl enzyme literature for a large-scale conformational change
upon substrate binding4 1,43'44. For methylcobalamin-dependent methionine synthase, the
enzyme exists as an ensemble of conformational states that interconvert upon substrate or
product binding4 4 . X-ray analysis reveals that the two active sites that alternatively
methylate
and demethylate
the Cbl cofactor
are -50
A
apart4 3 , requiring
large
conformational rearrangements of the enzyme during each catalytic cycle. For AdoCblbound MCM, the presence of substrate has a dramatic effect on the structure of the TIM
barrel41 . In the X-ray structure of the substrate-free form of MCM, the TIM barrel is
splayed open, leaving a large gap in the center of the barrel41 (Fig. 4.15c).
Substrate
binds in this gap, threading through the N-terminal end of the TIM barrel to the Cbl
cofactor, which is located at the C-terminal end of the barrel3 . Binding of substrate
appears to trigger the barrel closure (Fig. 4.15d), which in turn, causes Y89 to swing
toward the top of the AdoCbl corrin ring, presumably facilitating the homolytic cleavage
of the Ado moiety to give Ado.41 '45. Thus, while MCM and 5,6-LAM have similar
structures, the way in which conformational changes are linked to substrate binding is
different, and takes advantage of the unique properties of the substrates. The -23 A-long
methylmalonyl-CoA can bind along the entire length of the TIM barrel, causing the barrel
to tighten, whereas lysine, a much smaller substrate, can release the enzyme's imine
linkage from the PLP, freeing the Rossmann domain to rotate.
- 142-
4.4.2. Questions arisingfrom the structure of 5,6-LAM
Several open questions remain concerning the hypothetical conformational change that
we propose is necessary for catalysis in 5,6-LAM.
What role, if any, does the
dimerization domain play in the conformational change? Could the accessory clamp
swing up and down to facilitate the edge to top rotation of the Rossmann domain? What
drives the reformation of the imine linkage with K144P that would result in reversion to
the edge-on conformation and product release? What is the structural rationale for the
observed suicide inactivation of 5,6-LAM46 ? Although many interesting details remain to
be discovered, our structural analysis suggests a novel mechanism for substrate-mediated
control of radical generation. The structure of 5,6-LAM shows that AdoCbl-dependent
enzymes can control radical generation by using a covalent bond that must be broken
when substrates binds, effectively locking AdoCbl into a non-catalytic position in the
absence of substrate.
- 143-
4.5. Tables
Table 4.1. 5,6-LAM data collection and refinements statistics
Data Set
Wavelength
Se Peak
0.97918
50.0-2.8
45,836
(A)
Resolution range (A)
Unique Reflections 1
Redundancy
Completeness (%)
I/(I)
Rsym
Non-hydrogen atoms in asymmetric unit
Number of reflections (working set / test set)
6.8
99.8 (99.8)
18.6 (6.6)
9.3 (28.4)
5864
24098 / 2378
21.9 / 26.9
3
Rcryst
(%)/Rfree (%)
B (A2)
Overall
a (TIM barrel + accessory clamp)
main chain
31.3
36.1
side chain
3 (dimerization domain)
main chain
side chain
p (Rossmann-like domain)
main chain
side chain
Cofactors
Cbl
AdoH
62.9
67.7
31.6
38.1
34.8
70.1
24.9
PLP
RMS deviations of
protein bonds (A)
protein angles ()
Number of heavy atom sites in the asymmetric unit
Se
Co
0.009
1.4
29
1
Values in parentheses are for the highest-resolution shell. 1For the purpose of phasing,
Friedel pairs were not merged, and this is accounted for in the number of unique
reflections. 2 Rsy = [SkIs {Ii(hkl)- <I(hkl)>l] /kl
.I(hkl) for hkl independent
reflections and i observations of a given reflection. <I(hkl)> is the mean intensity of the
miller index (hkl).
3Rryst =
IF(hkl) - IFc(hkl)Il/
hkl IFo(hkl)I. Rfree= Rcrystfor a test
set of reflections (9.9% of all reflections) not included in refinement. No a cutoff was
used in the refinement.
- 144-
Table 4.2. Potential PLP-5,6-LAM hydrogen bonding and electrostatic
interactions
Atom
Contacting residue
Contacting atom
distance (A)
Pyridoxal
N1
03
03
Ser 238a
Asn 299cy
Asn 299a
OG
ND2
OD1
2.7
2.5
3.2
Phosphate
O1P
O1P
Arg 268a
Ser 189a
NE
OG
2.6
2.5
O1P
Tyr 263ca
OH
3.0
02P
02P
02P
02P
03P
03P
04P
Gln 188a
Ser 189a
Arg 184a
Ser 189a
Arg 268a
Gly 187a
Tyr 263a
N
N
NH1
OG
NH1
N
OH
3.3
3.0
2.5
3.1
2.7
2.5
2.9
- 145 -
Table 4.3. Potential AdoCbl-5,6-LAM hydrogen bonding interactions
Atom
Contacting residue
Contacting atom
distance (A)
03'
02'
Asp 54a
3.1
Glu 55ac
0
0
N6
Asp 64a
OD1
Thr 191f3
OG
N
N
N
AdoH
Corrin ring
N33
2.9
3.8
3.0
034
034
044
Thr 1913
N45
Thr 13003
O
051
051
Val 135/B
Thr 134/3
N
N
2.6
N52
Thr 134f/
OG1
3.5
R243f/
O
2.8
S187/3
OG
2.6
Gln 192/3
Ala 132f3
2.9
3.1
2.7
3.1
3.1
Ribose
08R
DMB
N3B
-146-
4.6. Figures
NH3
NH+3
H3 N
02
2 .-
Lysine
NH3+
3-LAM
12
AdoMet
H3N
5,6-LAM
AdoCbI
PLP
PLP
2,5-Diaminohexanoate
(2,5-DAH)
+~+02+
+-~'"'-~
'
CO 2
H3N
NH3
col
NH3
3,5-Diaminohexanoate
(3,5-DAH)
P-L-Lysine
Figure 4.1. Aminomutases in the bacterial lysine fermentation pathway. 5,6-LAM and
2,3-LAM catalyze similar reactions and act on similar substrates. Both enzymes require
PLP, but 5,6-LAM is AdoCbl dependent, whereas 2,3-LAM is an AdoMet-dependent
iron-sulfur enzyme. The natural substrates of 5,6-LAM include DL-Lysine and -LLysine. 2,3-LAM acts on L-Lysine and does not accept D-Lysine as a substrate.
- 147 -
(a)
Ado
,doCbl
AdoC
NH3
HC
NH3
t*LC
L-co
2
N CH
203PF
.
OH
NJs
.
.
Catalytic
"top-on"
conformaton
Cob(ll)
NH3
AdoH
H2 C" *2CO2
Cob(Ill)
AdoH
NNCH
2-
-
o~,o 'f
"
NH3
- - lHco2
H2C C
N
OH
+
Cob(ll)
N
NH3
AdoH
H2C' C"f
NCH
3
C ^02 /,,
PO
N
(b)
,
+rLys
11
NH3
O2C-TR
I
R
~N.-H
2-
0 3PO
N
,OH
OH
Figure 4.2. Proposed mechanism of 5,6-LAM. (a) Proposed mechanism of 5,6-LAM,
modified from reference4 2 . The boxed step represents the state of the enzyme observed in
our crystal structure. The unboxed steps are proposed to occur while 5,6-LAM is in the
hypothetical "top-on" conformation (see text).
catalyzed transaldimination with PLP47 -53.
- 148 -
(b) Mechanism of a general-base
(a)
aBr
COEt
2
-CH
6
(b)
N-
-
Bu3SnHIAIBN
-
t-BuOo
[. .
;INC NCH
<1300 C
Fr CO2Et
H2 C
Z CH
Nt CH
2I
6
>1300 C
N CH
CO2Et
N..CH
lw~
H2C1
2
N-C
N CH2
Figure. 4.3. (a) 1,2-imino shift under radical-generating conditions, as observed by Han
and Frey' 5 (b) Rearrangement of the aziridylcarbinyl radical, as observed by Danen and
West6 .
- 149-
-150-
Figure 4.4. Various crystal forms of 5,6-LAM. All crystals were produced using the
conventional hanging-drop vapor diffusion method (by mixing 1 L of protein solution
with 1 L of precipitant solution), and drops were equilibrated over a 0.5 mL well of
precipitant solution. (a) The form used to solve the structure (form 1). Only this crystal
form diffracts to give data of reasonable quality. Protein solution supplemented with
AdoCbl (12 mg/mL 5,6-LAM, 0.5 mM 3-mercaptoethanol,
10 mM triethanolamine pH
7.2, 4.5 mM AdoCbl; the molar ratio of AdoCbl to protein is 30) was mixed in a 1:1 ratio
with precipitant solution (0.1 M Tris hydrochloride pH 8.0, 0.2 M sodium acetate, 24%
w/v PEG 2000 monomethyl ether). Crystals appeared within 24 hours. (b) Protein
solution supplemented with AdoCbl (same as part a) was mixed in a 1:1 ratio with
precipitant solution (0.1 M sodium Cacodylate pH 6.5, 0.2 M sodium acetate, 24% w/v
PEG 2000 monomethyl ether). Crystals appeared within 24 hours. (c) Protein solution
supplemented with AdoCbl (same as part a) was mixed in a 1:1 ratio with precipitant
solution (0.1 M Tris hydrochloride pH 8.0, 0.2 M sodium acetate, 19% w/v PEG 2000
monomethyl ether, 5 mM NiCl2). Crystals appeared within 48 hours. (d) A solution of
protein, adeninylpentylcobalamin,
and substrate (12 mg/mL 5,6-LAM, 0.5 mM Pmercaptoethanol, 10 mM triethanolamine pH 7.2, 2 mM adeninylpentylcobalamin, 5 mM
D-lysine or L-lysine was mixed in a 1:1 ratio with precipitant solution (0.1 M Bis-tris pH
7.0, 2 M ammonium sulfate). The molar ratio of adeninylpentylcobalamin
to protein is
13. Crystals appeared within 7 days. (e) A solution of protein, CNCbl, and substrate (12
mg/mL 5,6-LAM, 0.5 mM -mercaptoethanol, 10 mM triethanolamine pH 7.2, 4.5 mM
CNCbl, 5 mM D-lysine or L-lysine was mixed in a 1:1 ratio with precipitant solution (0.1
M Bis-tris pH 6.5, 2 M ammonium sulfate). The molar ratio of CNCbl to protein is 30.
These crystals are small and only faintly pink, and appeared within 10 days. In this
picture, several are identified with arrows. Dark red crystals are probably CNCbl.
- 151 -
Figure 4.5. Solvent-flattened 3.0 A resolution experimental electron density of 5,6LAM. The map is contoured at 1.0 a. Residues K144f and N299,
shown.
- 152-
along with PLP, are
18) i
1
bi
135
a
_
I
I
-I
9E~()
a
E-
aIt
4
II
45 -
a AMU
a
(I:
a
.
)
a
I
I
l
i
I
i.
Co
I I I1
t
I
-45
a
tAN
I
p'
I
I
4
I35
-h
r-J
i
I
I
b
tP
ti
aa
UI
r
-b
II
!
135
a -0
-9)
45
.A5s
I
Phi (degrees S.>
Figure 4.6. Ramachandran plot of 5,6-LAM. The plot is colored as follows: red, most
favoured regions; yellow, additional allowed regions; light yellow, generously allowed
regions; white, disallowed regions. Glycine residues are shown as triangles. Figure
made using PROCHECK3 0 .
- 153-
Rossmann-like domain (1)
f
)
Figure 4.7. Overall structure of 5,6-LAM and location of cofactors. Ribbon diagram of
the a2 P2 5,6-LAM tetramer with the "accessory clamp" of a in brown, the TIM barrel of
a in green, the Rossmann-like domain and the dimerization domain of , in blue. The
dashed line represents the disordered loop connecting the two domains of . The
second ap unit is represented in darker colors of blue, green, and brown. AdoH is shown
in red sticks; Cbl in pink sticks and sphere; PLP in black sticks. With the exception of
Figs. 4.1, 4.2, and 4.3b, all figures were prepared using PyMOL
- 154-
54
.
I
I
I
I
Dimerization domain ()
Figure 4.8. Topology diagram of the 5,6-LAM ca4 heterodimer. The strands of the
TIM barrel and Rossmann-like domains are numbered in large black numerals. Red
numerals next to secondary structure elements indicate the N-terminal residue of that
element. The locations of H133P, which is the lower axial ligand to the cobalt of Cbl,
and K1443, which forms the imine bond to PLP, are emphasized with arrows. The Ado
binding loop, which forms most of the interactions between the protein and AdoH, is also
emphasized.
- 155 -
Figure 4.9. Lattice packing interactions of 5,6-LAM. The proposed physiologically
relevant 5,6-LAM tetramers in the crystal are shown in different colors. Each o unit
interacts with three other oe/ units of neighboring tetramers. PLP, Cbl, and AdoH are
shown in orange, pink, and red spheres, respectively.
-156-
121
(i
IrI
- --
(d)
,
/
(e)
-157-
I _
Figure 4.10. The five established PLP-binding enzyme families, plus 5,6-LAM. The
structures shown are those of (a) aspartate aminotransferase (PDB code 1AOG); (b)
tryptophan synthase (PDB code 1BKS); (c) alanine racemase (PDB code 1SFT); (d) Damino acid aminotransferase (PDB code 1DAA); (e) glycogen phosphorylase (PDB code
1FU4); (f) 5,6-LAM. The PLP binding subunits are colored in blue. PLP (black spheres)
and its associated Lys residue (black sticks) are shown. The PLP binding subunit of 5,6LAM (, the Rossmann-like domain) has a structure unlike that of any known PLP
binding protein.
- 158-
Figure 4.11. Superposition of the TIM barrels of alanine racemase and 5,6-LAM.
alanine racemase (pink) and 5,6-LAM (light green) are shown in cartoon representations.
The structures superimpose with an RMSD of 2.0 A2 for 66 accarbons, as determined by
LSQMAN55. The view is from the top of the TIM barrel. The PLP cofactor of alanine
racemase (red spheres) does not superimpose with the PLP of 5,6-LAM (dark green
spheres).
-159-
Figure 4.12. PLP in the putative active site of 5,6-LAM (a) Stereo view of the putative
active site of 5,6-LAM.
K1441 forms an imine bond to PLP; all other protein-PLP
interactions are made by residues of the TIM barrel. A simulated annealing composite
21Fol-IFcI
omit electron density map (orange mesh), contoured at 1.5 , is shown around
the PLP. Unless otherwise noted, the coloring scheme for all stick or ball-and-stick
representations is as follows: grey, C; red, O, blue, N; green, P; pink, Co. (b) Relative
positions of PLP and AdoCbl. PLP is inserted into the C-terminal end of the TIM barrel
by K144f3, which anchors the Rossmann-like domain in an off-center conformation on
the top comer of the TIM barrel. H133P, K1441, PLP, Cbl, and AdoH are all depicted in
black sticks. The secondary structural elements of the Rossmann-like domain that
contain H133P and K1443 are shown in a ribbon representation. Opaque domain
surfaces are shown, colored as in Fig. 4.7.
-160-
(a)
(b)
Y1
lp·-
Figure 4.13. The AdoCbl binding site of 5,6-LAM. (a) Stereo view of Cbl within a 1.8
omit electron density map, contoured at
A-radius simulated annealing composite 2IFol-lFcI
1.0 c. Hydrogen bonds between a propionamide side chain of the corrin ring and residues
T191
and Q192
are omitted for clarity. (b) Stereo view of AdoH. AdoCbl appears to
have cleaved in the X-ray beam, and we modeled the Ado moiety as AdoH. AdoH is
shown in a 1.8 A-radius simulated annealing 2Fol-IFcI omit electron density map
contoured at 1.0 a. Y193a is part of the TIM barrel domain; otherwise, all protein-AdoH
contacts are made by residues of the accessory clamp.
Figure 4.14. A stereo view of the local structure of the CxxCxxxC sequence. The three
Cys residues are not positioned for disulfide formation or metal ligation.
161 -
I.%I
I
l
won
a'-
.I
1
1951
(
Figure 4.15.
Hypothetical conformational change of 5,6-LAM and observed
conformational change of MCM. (a) Structure of the substrate-free form of 5,6-LAM, as
determined in this study. Protein domains and cofactors are colored as in Fig. 4.4a.
Arrows represent the axes of the TIM barrel and Rossmann-like domains. (b)
Hypothetical structure of the substrate-bound 5,6-LAM. The Rossmann-like domain is
modeled to cap the C-terminal end of the TIM barrel in a manner resembling the structure
of MCM. (c) Structure of the substrate-free form of MCM (PDB code 3REQ). The
Rossmann-like domain sits directly atop the center of the TIM barrel, which is pried open
in preparation for substrate binding. The Ado moiety of intact AdoCbl is shown in red,
and the Cbl portion is shown in pink. (d) Structure of substrate-bound MCM (PDB code
1REQ). The substrate fragment, desulfo-CoA (dark blue), threads through the TIM
barrel domain, effecting the closure of the TIM barrel to a much more compact structure.
The Ado moiety of AdoCbl was not observed.
- 162-
smannlomain (13)
barrel
(a)
Figure 4.16. A cleft leading to the putative active site of 5,6-LAM. Protein and cofactor
surfaces are shown, colored as in Fig. 4.4a, except that PLP is shown in orange. PLP is
entrenched in the active site. The cleft leading to the active site is formed by the
Rossmann-like domain and the dimerization domain.
- 163-
4.7. References
1.
2.
3.
Barker, H.A., Kahn, J.M. & Hedrick, L. Pathway of lysine degradation in
Fusobacterium nucleatum. J. Bact. 152, 201-207 (1982).
Reitzer, R. et al. Glutamate mutase from Clostridium cochlearium: the structure
of a coenzyme B12-dependent enzyme provides new mechanistic insights.
Structure Fold. Des. 7, 891-902 (1999).
Mancia, F. et al. How coenzyme B 12 radicals are generated: the crystal structure
of methylmalonyl-coenzyme A mutase at 2 A resolution. Structure 4, 339-350
(1996).
4.
Shibata, N. et al. A new mode of B12 binding and the direct participation of a
potassium ion in enzyme catalysis: X-ray structure of diol dehydratase. Structure
5.
Berkovtich, F., Nicolet, Y., Wan, J.T., Jarrett, J.T. & Drennan, C.L. Crystal
structure of biotin synthase, an S-adenosylmethionine-dependent radical enzyme.
7, 997-1008 (1999).
Science 303, 76-79 (2004).
6.
7.
8.
Berkovtich, F. et al. A locking mechanism preventing radical damage in the
absence of substrate, as revealed by the X-ray structure of lysine 5,6aminomutase. Proc. Natl. Acad. Sci., manuscript in press (2004).
Dekker, E.E. & Barker, H.A. Identification and Cobamide Coenzyme-dependent
Formation of 3,5-Diaminohexanoic Acid, and Intermediate in Lysine
Fermentation. J. Biol. Chem. 243, 3232-3237 (1968).
Morley, C.G.D. & Stadtman, T.C. Studies on the fermentation of D-a-Lysine.
Purification and properties of an adenosine triphosphate regulated B12 -coenzymedependent D-a-Lysine mutase complex from Clostridium sticklandii.
Biochemistry 9, 1970 (1970).
9.
10.
11.
12.
Chirpich, T.P., Zappia, V., Costilow, R.N. & Barker, H.A. Lysine 2,3aminomutase. Purification and properties of a pyridoxal phosphate and Sadenosylmethionine-activated enzyme. J. Biol. Chem. 245, 1778-1789 (1970).
Petrovich, R.M., Ruzicka, F.J., Reed, G.H. & Frey, P.A. Metal cofactors of
lysine-2,3-aminomutase. J. Biol. Chem. 266, 7656-7660 (1991).
Moss, M. & Frey, P.A. The role of S-adenosylmethionine in the lysine 2,3aminomutase reaction. J. Biol. Chem. 262, 14859-14862 (1987).
Morley, C.G.D. & Stadtman, T.C. The role of pyridoxal phosphate in the B1 2
coenzyme-dependent D-a-Lysine mutase reaction. Biochemistry 11, 600-605
(1972).
13.
14.
15.
Wetmore, S.D., Smith, D.M. & Random, L. How B6 Helps B12: The Roles of B6,
B12, and the Enzymes in Aminomutase-Catalyzed Reactions. J. Am. Chem. Soc.
122, 10208-10209 (2000).
Wetmore, S.D., Smith, D.M. & Random, L. Enzyme catalysis of 1,2-amino shifts:
the cooperative action of B6, B12, and aminomutases. J. Am. Chem. Soc. 123,
8678-8689 (2001).
Han, O. & Frey, P.A. Chemical model for the pyridoxal 5'-phosphate dependent
lysine aminomutases. J. Am. Chem. Soc. 112, 8982-8983 (1990).
- 164-
16.
17.
Danen, W.C. & West, C.T. An electron spin resonance investigation of the 1Aziridylcarbinyl and related free radicals. J. Am. Chem. Soc. 96, 2447-2453
(1973).
Zhao, Y., Such, P. & R6tey, J. Radical intermediates in the coenzyme B1 2
dependent methylmalonyl-CoA mutase reaction shown by ESR spectroscopy.
Angew. Chem. Int. Engl. 31, 215-216 (1992).
18.
19.
Marsh, E.N. & P., B.D. Coupling of cobalt-carbon bond homolysis and hydrogen
atom abstraction in adenosylcobalamin-dependent glutamate mutase.
Biochemistry 37, 11864-11872 (1998).
Zelder, O., Beatrix, B., Leutbecher, U. & Buckel, W. Characterization of the
coenzyme-B-12-dependent glutamate mutase from Clostridium-cochlearium
produced in Escherichia-coli. European Journal of Biochemistry 226, 577-585
(1994).
20.
Finlay, T.H., Valinsky, J., Mildvan, A.S. & Abeles, R.H. Electron spin resonance
studies with dioldehydrase. Evidence for radical intermediates in reactions
catalyzed by coenzyme B 12. J. Biol. Chem. 248, 1285-1290 (1973).
21.
Tamao, Y. & Blakley, R.L. Direct spectroscopic observation of an intermediate
formed from deoxyadenosylcoblamain in ribonucleotide reductase. Biochemistry
12, 24-34 (1973).
22.
23.
24.
25.
26.
27.
28.
29.
30.
Budisa, N. et al. High-level biosynthetic substitution of methionine in proteins by
its analogs 2-aminohexanoic acid, selenomethionine, telluromethionine and
ethionine in Escherichia coli. Eur. J. Biochem. 230, 788-796 (1995).
Chang, C.H. & Frey, P.A. Cloning, sequencing, heterologous expression,
purification, and characterization of adenosylcobalamin-dependent D-lysine 5, 6aminomutase from Clostridium sticklandii. J. Biol. Chem. 275, 106-114 (2000).
Otwinowski, Z. & Minor, W. Processing of X-ray diffraction data collected in
oscillation mode. Methods Enzymol. 276, 307-326 (1997).
Rice, L.M., Earnest, T.N. & Brunger, A.T. Single-wavelength anomalous
diffraction phasing revisited. Acta. Cryst. D 56, 1413-1420 (2000).
Dodson, E. Is it jolly SAD? Acta. Cryst. D 59, 1958-1965 (2003).
Terwilliger, T.C. & Berendzen, J. Automated structure solution for MIR and
MAD. Acta. Cryst. D55, 849-861 (1999).
McRee, D.E. XtalView/Xfit--A versatile program for manipulating atomic
coordinates and electron density. J. Struct. Biol. 123, 156-165 (1999).
Brunger, A.T. et al. Crystallography & NMR system: A new software suite for
macromolecular structure determination. Acta. Cryst. D55, 905-921 (1998).
Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton, J.M. PROCHECK:
a program to check the stereochemical quality of protein structures. J. Appl. Cryst.
26, 283-291 (1993).
31.
Baker, J.J., van der Drift, C. & Stadtman, T.C. Purification and properties of Plysine mutase, a pyridoxal phosphate and B12 coenzyme dependent enzyme.
Biochemistry 12, 1054-1063 (1973).
32.
Barnham, K.J. et al. Structure of the Alzheimer's disease amyloid precursor
protein copper binding domain. A regulator of neuronal copper homeostasis. J.
Biol. Chem. 278, 17401-17407 (2003).
165 -
33.
Schneider, G., Kack, H. & Lindqvist, Y. The manifold of vitamin B6 dependent
34.
enzymes. Structure Fold. Des. 8, R1-R6 (2000).
Tang, K.H., Harms, A. & Frey, P.A. Identification of a novel pyridoxal 5'-
35.
36.
phosphate binding site in adenosylcobalamin-dependent lysine 5,6-aminomutase
from Porphyromonas gingivalis. Biochemistry 41, 8767-8776 (2002).
Hester, G. et al. Crystal structure of phosphoserine aminotransferase from
Escherichia coli at 2.3 A resolution: comparison of the unligated enzyme and a
complex with alpha-methyl-l-glutamate. J. Mol. Biol. 286, 829-850 (1999).
Weyand, M. & Schlichting, I. Crystal structure of wild-type tryptophan synthase
complexed with the natural substrate indole-3-glycerol phosphate. Biochemistry
38, 16469-16480 (1999).
37.
38.
39.
Tirupati, B., Vey, J.L., Drennan, C.L. & Bollinger, J.M., Jr. Kinetic and structural
characterization of Slr0077/SufS, the essential cysteine desulfurase from
Synechocystis sp. PCC 6803. Biochemistry In press(2004).
Drennan, C.L., Huang, S., Drummond, J.T., Matthews, R.G. & Ludwig, M.L.
How a protein binds B12: A 3.0 A X-ray structure of B 12 -binding domains of
methionine synthase. Science 266, 1669-1674 (1994).
Champloy, F., Gruber, K., Jogl, G. & Kratky, G. XAS spectroscopy reveals X-
ray-induced photoreduction of free and protein-bound B-12 cofactors. J. Synch.
Rad. 7, 267-273 (2000).
40.
Gruber, K., Reitzer, R. & Kratky, G. Radical Shuttling in a Protein: Ribose
Pseudorotation Controls Alkyl-Radical Transfer in the Coenzyme B(12)
Dependent Enzyme Glutamate Mutase. Angew. Chem. Int. Ed. 40, 3377-3380
(2001).
41.
42.
43.
44.
45.
46.
Mancia, F. & Evans, P.R. Conformational changes on substrate binding to
methylmalonyl CoA mutase and new insights into the free radical mechanism.
Structure 6, 711-720 (1998).
Tang, K.H., Casarez, A.D., Wu, W. & Frey, P.A. Kinetic and biochemical
analysis of the mechanism of action of lysine 5,6-aminomutase. Arch. Biochem.
Biophys. 418, 49-54 (2003).
Evans, J.C. et al. Structures of the N-terminal modules imply large domain
motions during catalysis by methionine synthase. Proc. Natl. Acad. Sci. 101,
3729-3726 (2004).
Bandarian, V., Ludwig, M.L. & Matthews, R.G. Factors modulating
conformational equilibria in large modular proteins: a case study with cobalamindependent methionine synthase. Proc. Natl. Acad. Sci. 100, 8156-8163 (2003).
Vlasie, M.D. & Banerjee, R. Tyrosine 89 accelerates Co-carbon bond homolysis
in methylmalonyl-CoA mutase. J. Am. Chem. Soc. 125, 5431-5435 (2003).
Tang, K.H., Chang, C.H. & Frey, P.A. Electron transfer in the substrate-
dependent suicide inactivation of Lysine 5,6-Aminomutase. Biochemistry 40,
5190-5199 (2001).
47.
Auld, D.S. & Bruice, T.C. Catalytic reactions involving azomethines. VII. Rates
and equilibria of aldimine formation with 3-hydroxypyridine-4-aldehyde and
alanine. J. Am. Chem. Soc. 89, 2083-2089 (1967).
- 166-
48.
Auld, D.S. & Bruice, T.C. Catalytic reactions involving azomethines. VIII. Water
and alanine catalysis of the transamination of 3-hydroxypyridine-4-aldehyde by
alanine. J. Am. Chem. Soc. 89, 2090-2097 (1967).
49.
Auld, D.S. & Bruice, T.C. Catalytic reactions involving azomethines. IX. General
base catalysis of the transamination of 3-hydroxypyridine-4-aldehyde by alanine.
J. Am. Chem. Soc. 89, 2098-2106 (1967).
50.
Harris, S.A., Webb, T.J. & Folkers, K. Chemistry of vitamin B6. I. Tautomerism.
J. Am. Chem. Soc. 62, 3198-3203 (1940).
51.
Metzler, D.E. Equilibria between pyridoxal and amino acids and their imines. J.
Am. Chem. Soc. 79, 485-490 (1957).
52.
Metzler, D.E. & Snell, E.E. Spectra and ionization constants of the vitamin B6
group and related 3-hydroxypyridine derivatives. J. Am. Chem. Soc. 77, 24312437 (1955).
53.
Thanassi, J.W., Butler, A.R. & Bruice, T.C. Catalytic reactions involving
azomethines. VI. The mechanism of the transamination of 3-hydroxypyridine-4-
54.
DeLano, W.L. The PyMOL Molecular Graphics System.
55.
Kleywegt, G.J. Use of non-crystallographic symmetry in protein structure
refinement. Acta. Cryst. D52, 842-857 (1996).
aldehyde by glutamic acid. Biochemistry 4, 1463-1472 (1965).
( ttp: /vmol I.source folre. net).
- 167 -
Chapter 5
Conclusions
- 165-
Abbreviations:
2,3-LAM
5,6-LAM
Ado
Ado*
AdoCbl
AdoH
AdoMet
BioB
Cbl
CNCbl
DDH
DTB
HemN
RNR
SAM
Lysine 2,3-aminomutase
Lysine 5,6-aminomutase
5'-deoxyadenosyl group
5'-deoxyadenosyl radical
Adenosylcobalamin, Coenzyme B 12
5'-deoxyadenosine
S-adenosylmethionine
Biotin synthase
Cobalamin
Cyanocobalamin, Vitamin B12
Diol dehydratase
Dethiobiotin
Coproporphyrinogen III oxidase
Ribonucleotide reductase
S-adenosylmethionine
- 166 -
The crystal structures of biotin synthase (BioB) and lysine 5,6-aminomutase (5,6-LAM)
allow for a comparison of the adenosylcobalamin (AdoCbl) and the adenosylmethionine
(AdoMet) -dependent radical enzymes. In particular, we learn that at least one AdoMet
radical enzyme has a TIM barrel fold', like most of the Cbl-dependent enzymes2 '6 . In a
break from the TIM barrel trend, the structure of HemN, the only other AdoMet radical
enzyme structure that is published, has an a/fl fold (referred to as a "3/4 barrel") that is
reminiscent of the TIM barrel 7 .
The elucidation of the structure of 2,3-LAM is currently under way (P. A. Frey, personal
communication). It is not known if its structure resembles the 3/4 barrel of HemN, the
TIM barrel of BioB and 5,6-LAM, or some other structure.
5.1 Structural comparison of BioB with HemN
Superposition of BioB with HemN reveals that the two enzymes share a core set of strands and a helices (Fig. 5.1a). The common set of strands and helices superimpose
with a root mean square deviation of 1.9
2
for 99 alpha carbons (Fig. 5.lb).
Additionally, the 4Fe-4S cluster, its coordinating AdoMet, and several important residues
are structurally equivalent (Fig. 5.1c). In BioB, these residues are: Y59 and 1192, which
interact with the adenine ring of AdoMet; D155, which hydrogen bonds to the 2' and 3' OH groups of the ribose moiety; and R173, which forms a salt bridge with the carboxyl
group of AdoMet. Interestingly, a second AdoMet molecule, proposed to give rise to the
second Adoe in the hypothetical HemN mechanism 7, was modeled into the structure of
- 167 -
HemN. The second AdoMet molecule is located near the DTB site of BioB, possibly
suggesting that this AdoMet might be occupying the binding site of the heme substrate.
5.2 Structural comparison of BioB with diol dehydratase
One tentative indication of an evolutionary link between the AdoMet and the AdoCbl
radical enzymes is evident upon structural comparison of BioB and diol dehydratase
(DDH). Superposition of the alpha carbons of both crystal structures (DDH PDB code
1DI0 5) reveals several remarkable similarities (Fig. 5.2).
The Co atom of
cyanocobalamin (CNCbl) occupies the same space as the 4Fe-4S cluster of BioB, a
catalytic K+ ion in DDH coincides with the 2Fe-2S cluster of BioB, the substrates of both
enzymes superimpose, and both substrates are covered by similar loops.
Clearly, BioB, as well as the AdoCbl-dependent enzymes (with the exception of the class
II RNR), are able to properly arrange their catalytic elements using the TIM barrel fold.
It is possible that the TIM barrel is especially adept at catalyzing Ado. chemistry because
the radical intermediates can be effectively sequestered inside the barrel.
The structure of BioB is more like the structure of DDH than that of HemN. Both BioB
and DDH have relatively
small substrates,
compared
to the HemN substrate,
a
tetrapyrrole macrocycle. Accordingly, the 3/4 barrel fold of HemN envelopes more
volume than does a TIM barrel, allowing for a larger active site. It is tempting to
speculate that substrate size dictates the use of a 3/4 barrel fold over a TIM barrel; a 3/4
barrel would be required for large substrates which could not fit inside an eight-stranded
- 168-
fl-barrel structure. With the current limitations on the number of AdoMet radical enzyme
structures, there is no convincing support for this theory. To this end, the structure of
2,3-LAM is of special interest, since its substrate is identical to one of the 5,6-LAM
substrates. If substrate size does determine the fold of an Ado. enzyme, then we expect
2,3-LAM to have a TIM barrel fold, like 5,6-LAM.
5.3. A novel role for PLP in an Adoe enzyme
We propose that in 5,6-LAM, PLP has a novel role in keeping AdoCbl out of the active
site in the absence of substrate. Transimination of the covalent PLP-K144/gimine linkage
with the substrate releases the Rossmann domain from its constrained position, allowing
it to rotate and place AdoCbl into the active site of the TIM barrel domain. To our
knowledge, this is the first reported instance of PLP being used to lock an enzyme in a
certain conformation.
5.4. Ado. enzymes: barrel structures with reaction-specific accommodations
Ado* enzymes commonly utilize barrel-like structures, which are adapted to the
particular enzymatic reaction. BioB appears to be very similar in structure to AdoCbl
enzymes, but the AdoCbl-binding domain is replaced by a loop at the top of the barrel
that binds the essential 4Fe-4S cluster. The structure of 5,6-LAM is generally similar to
those of other AdoCbl enzymes, in that it contains a TIM barrel and a Rossmann-like
domain. However, these domains are modified to create a mechanism by which a small
substrate, lysine, can trigger a large conformational change upon binding and
transimination.
A broken helix of the Rossmann-like domain creates a novel PLP
-169-
binding motif.
This motif allows for PLP-mediated interactions to control the
conformation of the enzyme, keeping AdoCbl out of the active site in the absence of
substrate. Thus, we expect to see more variation on the general theme of aS/ barrel
structures as more AdoMet and AdoCbl radical enzymes are structurally characterized.
-170-
5.5. Figures
-
i
(a)
I
%
I^%
Figure 5.1. (a) Superposition of BioB (blue) and HemN (grey). The view is from the
bottom of the BioB TIM barrel. Cofactors and substrates are omitted for clarity. (b)
Stereo view alpha carbon trace of the structurally homologous regions of BioB and
HemN.
The color scheme for parts b and c are the same as for part a. (c) Analogous
residues in the active sites of BioB and HemN (see text). Residues are labeled according
to the coloring scheme of part a. The 4Fe-4S clusters and associated AdoMet of both
proteins are labeled in black typeface. The second AdoMet of HemN is labeled
"AdoMet2" in grey. The 2Fe-2S cluster of BioB is labeled "2Fe" in blue. Figures 5.1
and 5.2 were made using PyMOL 8 .
171 -
References
1.
Berkovtich, F., Nicolet, Y., Wan, J.T., Jarrett, J.T. & Drennan, C.L. Crystal
structure of biotin synthase, an S-adenosylmethionine-dependent radical enzyme.
Science 303, 76-79 (2004).
2.
3.
Berkovtich, F. et al. A locking mechanism preventing radical damage in the
absence of substrate, as revealed by the X-ray structure of lysine 5,6aminomutase. Proc. Natl. Acad. Sci., manuscript in press (2004).
Drennan, C.L., Huang, S., Drummond, J.T., Matthews, R.G. & Ludwig, M.L.
How a protein binds B12: A 3.0 A X-ray structure of B 12 -binding domains of
methionine synthase. Science 266, 1669-1674 (1994).
4.
Mancia, F. et al. How coenzyme B12 radicals are generated: the crystal structure
of methylmalonyl-coenzyme A mutase at 2 A resolution. Structure 4, 339-350
(1996).
5.
Shibata, N. et al. A new mode of B 12 binding and the direct participation of a
potassium ion in enzyme catalysis: X-ray structure of diol dehydratase. Structure
7, 997-1008 (1999).
6.
7.
8.
Reitzer, R. et al. Glutamate mutase from Clostridium cochlearium: the structure
of a coenzyme B 12-dependent enzyme provides new mechanistic insights.
Structure Fold. Des. 7, 891-902 (1999).
Layer, G., Moser, J., Heinz, D.W., Jahn, D. & Schubert, W.D. Crystal structure of
coproporphyrinogen III oxidase reveals cofactor geometry of Radical SAM
enzymes. EMBO J. 22, 6214-6224 (2003).
DeLano, W.L. The PyMOL Molecular Graphics System.
(http:/'/pymol.sourceforge.net).
- 173-
Appendices
Appendix 1. Summary of 5,6-LAM crystallization experiments. Commercially
available screens (Hampton Research Corp.) were used (denoted by "yes") or not used
(denoted by "no") in crystallization experiments with 5,6-LAM, AdoCbl,
cyanocobalamin (CNCbl), Adeninylpentylcobalamin (APCbl), substrate (D or L-lysine),
or inhibitors (L-ornithine [L-orn] or thialysine [Thialys]), in various combinations. For
details of crystallization experiments, see chapter 4.
Crystal
screen
lite
Index Index Memb-
AS
PEG/Ion
I
II
Fac
grid
grid
PEG
6000
grid
Natrix
screen 2
5,6-LAM
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
5,6-LAM
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
Yes
Yes
No
Yes
Yes
Yes
No
Crystal
Protein
sample
+AdoCbl
5,6-LAM
+AdoCbl
+Thialys
5,6-LAM
+AdoCbl
+L-Orn
5,6-LAM
+APCbl
+ D-Lys
5,6-LAM
+CNCbl
+ D-Lys
5,6-LAM
+CNCbl
+ L-Lys
- 174-
Room 14-0551
MITLibraries
Document Services
77 Massachusetts Avenue
Cambridge, MA 02139
Ph: 617.253.5668 Fax: 617.253.1690
Email: docs@mit.edu
http://libraries.mit.edu/docs
DISCLAIMER OF QUALITY
Due to the condition of the original material, there are unavoidable
flaws in this reproduction. We have made every effort possible to
provide you with the best copy available. If you are dissatisfied with
this product and find it unusable, please contact Document Services as
soon as possible.
Thank you.
PAGES 165, 166, 167 APPEAR TWICE.